/[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 87 by nigel, Sat Feb 24 21:41:21 2007 UTC revision 389 by ph10, Sun Mar 15 18:24:05 2009 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2006 University of Cambridge             Copyright (c) 1997-2009 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 54  functions whose names all begin with "_p Line 54  functions whose names all begin with "_p
54  /* 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
55  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
56  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
57  all, it had only been about 10 years then... */  all, it had only been about 10 years then...
58    
59    It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so
60    be absolutely sure we get our version. */
61    
62    #undef DPRINTF
63  #ifdef DEBUG  #ifdef DEBUG
64  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
65  #else  #else
66  #define DPRINTF(p) /*nothing*/  #define DPRINTF(p) /* Nothing */
67  #endif  #endif
68    
69    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
70  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
71  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
72    
# Line 79  setjmp and stdarg are used is when NO_RE Line 79  setjmp and stdarg are used is when NO_RE
79  #include <stdlib.h>  #include <stdlib.h>
80  #include <string.h>  #include <string.h>
81    
82  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
83  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
84    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
85    information there, using __declspec(dllexport) without "extern" we have a
86    definition; with "extern" we have a declaration. The settings here override the
87    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
88    which is all that is needed for applications (they just import the symbols). We
89    use:
90    
91      PCRE_EXP_DECL       for declarations
92      PCRE_EXP_DEFN       for definitions of exported functions
93      PCRE_EXP_DATA_DEFN  for definitions of exported variables
94    
95    The reason for the two DEFN macros is that in non-Windows environments, one
96    does not want to have "extern" before variable definitions because it leads to
97    compiler warnings. So we distinguish between functions and variables. In
98    Windows, the two should always be the same.
99    
100    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
101    which is an application, but needs to import this file in order to "peek" at
102    internals, can #include pcre.h first to get an application's-eye view.
103    
104    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
105    special-purpose environments) might want to stick other stuff in front of
106    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
107    PCRE_EXP_DATA_DEFN only if they are not already set. */
108    
109    #ifndef PCRE_EXP_DECL
110    #  ifdef _WIN32
111    #    ifndef PCRE_STATIC
112    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
113    #      define PCRE_EXP_DEFN       __declspec(dllexport)
114    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
115    #    else
116    #      define PCRE_EXP_DECL       extern
117    #      define PCRE_EXP_DEFN
118    #      define PCRE_EXP_DATA_DEFN
119    #    endif
120    #  else
121    #    ifdef __cplusplus
122    #      define PCRE_EXP_DECL       extern "C"
123    #    else
124    #      define PCRE_EXP_DECL       extern
125    #    endif
126    #    ifndef PCRE_EXP_DEFN
127    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
128    #    endif
129    #    ifndef PCRE_EXP_DATA_DEFN
130    #      define PCRE_EXP_DATA_DEFN
131    #    endif
132    #  endif
133    #endif
134    
135    /* When compiling with the MSVC compiler, it is sometimes necessary to include
136    a "calling convention" before exported function names. (This is secondhand
137    information; I know nothing about MSVC myself). For example, something like
138    
139      void __cdecl function(....)
140    
141    might be needed. In order so make this easy, all the exported functions have
142    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
143    set, we ensure here that it has no effect. */
144    
145    #ifndef PCRE_CALL_CONVENTION
146    #define PCRE_CALL_CONVENTION
147  #endif  #endif
148    
149  /* 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 154  preprocessor time in standard C environm
154    
155  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
156    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
157      typedef short pcre_int16;
158  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
159    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
160      typedef int pcre_int16;
161  #else  #else
162    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
163  #endif  #endif
164    
165  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
166    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
167      typedef int pcre_int32;
168  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
169    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
170      typedef long int pcre_int32;
171  #else  #else
172    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
173  #endif  #endif
# Line 114  Unix, where it is defined in sys/types, Line 181  Unix, where it is defined in sys/types,
181    
182  typedef unsigned char uschar;  typedef unsigned char uschar;
183    
184    /* This is an unsigned int value that no character can ever have. UTF-8
185    characters only go up to 0x7fffffff (though Unicode doesn't go beyond
186    0x0010ffff). */
187    
188    #define NOTACHAR 0xffffffff
189    
190    /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
191    "any" and "anycrlf" at present). The following macros are used to package up
192    testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
193    modules to indicate in which datablock the parameters exist, and what the
194    start/end of string field names are. */
195    
196    #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
197    #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
198    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
199    
200    /* This macro checks for a newline at the given position */
201    
202    #define IS_NEWLINE(p) \
203      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
204        ((p) < NLBLOCK->PSEND && \
205         _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
206           utf8)) \
207        : \
208        ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
209         (p)[0] == NLBLOCK->nl[0] && \
210         (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
211        ) \
212      )
213    
214    /* This macro checks for a newline immediately preceding the given position */
215    
216    #define WAS_NEWLINE(p) \
217      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
218        ((p) > NLBLOCK->PSSTART && \
219         _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
220           &(NLBLOCK->nllen), utf8)) \
221        : \
222        ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
223         (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
224         (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
225        ) \
226      )
227    
228  /* When PCRE is compiled as a C++ library, the subject pointer can be replaced  /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
229  with a custom type. This makes it possible, for example, to allow pcre_exec()  with a custom type. This makes it possible, for example, to allow pcre_exec()
230  to process subject strings that are discontinuous by using a smart pointer  to process subject strings that are discontinuous by using a smart pointer
# Line 131  must begin with PCRE_. */ Line 242  must begin with PCRE_. */
242  #define USPTR const unsigned char *  #define USPTR const unsigned char *
243  #endif  #endif
244    
245    
246    
247  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
248  values. */  values. */
249    
# Line 142  need to have their names changed. PCRE m Line 255  need to have their names changed. PCRE m
255  option on the command line. */  option on the command line. */
256    
257  #ifdef VPCOMPAT  #ifdef VPCOMPAT
258    #define strlen(s)        _strlen(s)
259  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
260    #define memcmp(s,c,n)    _memcmp(s,c,n)
261  #define memcpy(d,s,n)    _memcpy(d,s,n)  #define memcpy(d,s,n)    _memcpy(d,s,n)
262  #define memmove(d,s,n)   _memmove(d,s,n)  #define memmove(d,s,n)   _memmove(d,s,n)
263  #define memset(s,c,n)    _memset(s,c,n)  #define memset(s,c,n)    _memset(s,c,n)
# Line 151  option on the command line. */ Line 266  option on the command line. */
266  /* 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(),
267  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
268  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
269  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. */  
270    
271  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
272  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
273  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
274  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
275  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
276  void *  static void *
277  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
278  {  {
279  size_t i;  size_t i;
280  dest += n;  unsigned char *dest = (unsigned char *)d;
281  src += n;  const unsigned char *src = (const unsigned char *)s;
282  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
283  return dest;    {
284      dest += n;
285      src += n;
286      for (i = 0; i < n; ++i) *(--dest) = *(--src);
287      return (void *)dest;
288      }
289    else
290      {
291      for (i = 0; i < n; ++i) *dest++ = *src++;
292      return (void *)(dest - n);
293      }
294  }  }
295  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
296  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 253  capturing parenthesis numbers in back re Line 376  capturing parenthesis numbers in back re
376    
377  /* 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
378  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
379  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
380    never be called in byte mode. To make sure it can never even appear when UTF-8
381    support is omitted, we don't even define it. */
382    
383  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
384  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 261  byte-mode, and more complicated ones for Line 386  byte-mode, and more complicated ones for
386  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
387  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
388  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
389  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
390    
391  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
392    
# Line 270  we know we are in UTF-8 mode. */ Line 395  we know we are in UTF-8 mode. */
395    
396  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
397    c = *eptr; \    c = *eptr; \
398    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
399      { \      { \
400      int gcii; \      int gcii; \
401      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 288  pointer. */ Line 413  pointer. */
413    
414  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
415    c = *eptr; \    c = *eptr; \
416    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
417      { \      { \
418      int gcii; \      int gcii; \
419      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 306  know we are in UTF-8 mode. */ Line 431  know we are in UTF-8 mode. */
431    
432  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
433    c = *eptr++; \    c = *eptr++; \
434    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
435      { \      { \
436      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
437      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 322  know we are in UTF-8 mode. */ Line 447  know we are in UTF-8 mode. */
447    
448  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
449    c = *eptr++; \    c = *eptr++; \
450    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
451      { \      { \
452      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
453      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 339  if there are extra bytes. This is called Line 464  if there are extra bytes. This is called
464    
465  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
466    c = *eptr; \    c = *eptr; \
467    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
468        { \
469        int gcii; \
470        int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
471        int gcss = 6*gcaa; \
472        c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
473        for (gcii = 1; gcii <= gcaa; gcii++) \
474          { \
475          gcss -= 6; \
476          c |= (eptr[gcii] & 0x3f) << gcss; \
477          } \
478        len += gcaa; \
479        }
480    
481    /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
482    pointer, incrementing length if there are extra bytes. This is called when we
483    know we are in UTF-8 mode. */
484    
485    #define GETCHARLENTEST(c, eptr, len) \
486      c = *eptr; \
487      if (utf8 && c >= 0xc0) \
488      { \      { \
489      int gcii; \      int gcii; \
490      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 354  if there are extra bytes. This is called Line 499  if there are extra bytes. This is called
499      }      }
500    
501  /* 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
502  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
503    because almost all calls are already within a block of UTF-8 only code. */
504    
505  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
506    
507  #endif  #endif
508    
# Line 373  Standard C system should have one. */ Line 519  Standard C system should have one. */
519    
520  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
521    
522  /* Private options flags start at the most significant end of the four bytes,  /* Private flags containing information about the compiled regex. They used to
523  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
524  with negative values. The public options defined in pcre.h start at the least  are in a 16-bit flags word. */
525  significant end. Make sure they don't overlap! */  
526    #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
527  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
528  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
529  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
530  #define PCRE_ICHANGED      0x08000000  /* i option changes within regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
531  #define PCRE_NOPARTIAL     0x04000000  /* can't use partial with this regex */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
532    
533  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
534    
# Line 391  significant end. Make sure they don't ov Line 537  significant end. Make sure they don't ov
537  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
538  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
539    
540  #define PUBLIC_OPTIONS \  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
541                               PCRE_NEWLINE_ANYCRLF)
542    
543    #define PUBLIC_COMPILE_OPTIONS \
544    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
545     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
546     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| \
547       PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
548       PCRE_JAVASCRIPT_COMPAT)
549    
550  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
551    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
552     PCRE_PARTIAL)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
553       PCRE_NO_START_OPTIMIZE)
554    
555  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
556    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
557     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
558       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
559    
560  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
561    
# Line 427  variable-length repeat, or a anything ot Line 580  variable-length repeat, or a anything ot
580  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
581  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
582    
583  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
584    environments where these macros are defined elsewhere. Unfortunately, there
585    is no way to do the same for the typedef. */
586    
587  typedef int BOOL;  typedef int BOOL;
588    
589    #ifndef FALSE
590  #define FALSE   0  #define FALSE   0
591  #define TRUE    1  #define TRUE    1
592    #endif
593    
594  /* Escape items that are just an encoding of a particular data value. Note that  /* Escape items that are just an encoding of a particular data value. */
 ESC_n is defined as yet another macro, which is set in config.h to either \n  
 (the default) or \r (which some people want). */  
595    
596  #ifndef ESC_e  #ifndef ESC_e
597  #define ESC_e 27  #define ESC_e 27
# Line 447  ESC_n is defined as yet another macro, w Line 602  ESC_n is defined as yet another macro, w
602  #endif  #endif
603    
604  #ifndef ESC_n  #ifndef ESC_n
605  #define ESC_n NEWLINE  #define ESC_n '\n'
606  #endif  #endif
607    
608  #ifndef ESC_r  #ifndef ESC_r
# Line 485  contain UTF-8 characters with values gre Line 640  contain UTF-8 characters with values gre
640  value such as \n. They must have non-zero values, as check_escape() returns  value such as \n. They must have non-zero values, as check_escape() returns
641  their negation. Also, they must appear in the same order as in the opcode  their negation. Also, they must appear in the same order as in the opcode
642  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  definitions below, up to ESC_z. There's a dummy for OP_ANY because it
643  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
644  ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two  (which is used for [^] in JavaScript compatibility mode).
 tests in the code for an escape greater than ESC_b and less than ESC_Z to  
 detect the types that may be repeated. These are the types that consume  
 characters. If any new escapes are put in between that don't consume a  
 character, that code will have to change. */  
   
 enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  
        ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E,  
        ESC_Q, ESC_REF };  
645    
646  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  The final escape must be ESC_REF as subsequent values are used for
647  that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
648    greater than ESC_b and less than ESC_Z to detect the types that may be
649    repeated. These are the types that consume characters. If any new escapes are
650    put in between that don't consume a character, that code will have to change.
651    */
652    
653    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
654           ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
655           ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
656           ESC_REF };
657    
658    
659    /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
660  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.
661  Note that whenever this list is updated, the two macro definitions that follow  
662  must also be updated to match. */  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
663    that follow must also be updated to match. There is also a table called
664    "coptable" in pcre_dfa_exec.c that must be updated. */
665    
666  enum {  enum {
667    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 509  enum { Line 670  enum {
670    
671    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
672    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
673    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
674    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
675    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
676    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
677    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
678    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
679    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
680    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
681    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
682    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
683    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
684    OP_PROP,           /* 14 \p (Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
685    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
686    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_PROP,           /* 16 \p (Unicode property) */
687    OP_EOD,            /* 17 End of data: \z */    OP_ANYNL,          /* 17 \R (any newline sequence) */
688      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
689    OP_OPT,            /* 18 Set runtime options */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
690    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
691    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
692    OP_CHAR,           /* 21 Match one character, casefully */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
693    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
694    OP_NOT,            /* 23 Match anything but the following char */    OP_EOD,            /* 24 End of data: \z */
695    
696    OP_STAR,           /* 24 The maximizing and minimizing versions of */    OP_OPT,            /* 25 Set runtime options */
697    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
698    OP_PLUS,           /* 26 the minimizing one second. */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
699    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_CHAR,           /* 28 Match one character, casefully */
700    OP_QUERY,          /* 28 */    OP_CHARNC,         /* 29 Match one character, caselessly */
701    OP_MINQUERY,       /* 29 */    OP_NOT,            /* 30 Match one character, not the following one */
702    OP_UPTO,           /* 30 From 0 to n matches */  
703    OP_MINUPTO,        /* 31 */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
704    OP_EXACT,          /* 32 Exactly n matches */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
705      OP_PLUS,           /* 33 the minimizing one second. */
706    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
707    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */    OP_QUERY,          /* 35 */
708    OP_NOTPLUS,        /* 35 the minimizing one second. */    OP_MINQUERY,       /* 36 */
709    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */  
710    OP_NOTQUERY,       /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
711    OP_NOTMINQUERY,    /* 38 */    OP_MINUPTO,        /* 38 */
712    OP_NOTUPTO,        /* 39 From 0 to n matches */    OP_EXACT,          /* 39 Exactly n matches */
713    OP_NOTMINUPTO,     /* 40 */  
714    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_POSSTAR,        /* 40 Possessified star */
715      OP_POSPLUS,        /* 41 Possessified plus */
716    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_POSQUERY,       /* 42 Posesssified query */
717    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */    OP_POSUPTO,        /* 43 Possessified upto */
718    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */  
719    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
720    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
721    OP_TYPEMINQUERY,   /* 47 */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
722    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
723    OP_TYPEMINUPTO,    /* 49 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
724    OP_TYPEEXACT,      /* 50 Exactly n matches */    OP_NOTMINQUERY,    /* 49 */
725    
726    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */    OP_NOTUPTO,        /* 50 From 0 to n matches */
727    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    OP_NOTMINUPTO,     /* 51 */
728    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */    OP_NOTEXACT,       /* 52 Exactly n matches */
729    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */  
730    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
731    OP_CRMINQUERY,     /* 56 */    OP_NOTPOSPLUS,     /* 54 */
732    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_NOTPOSQUERY,    /* 55 */
733    OP_CRMINRANGE,     /* 58 */    OP_NOTPOSUPTO,     /* 56 */
734    
735      OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
736      OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
737      OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
738      OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
739      OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
740      OP_TYPEMINQUERY,   /* 62 */
741    
742      OP_TYPEUPTO,       /* 63 From 0 to n matches */
743      OP_TYPEMINUPTO,    /* 64 */
744      OP_TYPEEXACT,      /* 65 Exactly n matches */
745    
746      OP_TYPEPOSSTAR,    /* 66 Possessified versions */
747      OP_TYPEPOSPLUS,    /* 67 */
748      OP_TYPEPOSQUERY,   /* 68 */
749      OP_TYPEPOSUPTO,    /* 69 */
750    
751      OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
752      OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
753      OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
754      OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
755      OP_CRQUERY,        /* 74 These are for character classes and back refs */
756      OP_CRMINQUERY,     /* 75 */
757      OP_CRRANGE,        /* 76 These are different to the three sets above. */
758      OP_CRMINRANGE,     /* 77 */
759    
760    OP_CLASS,          /* 59 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
761    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
762                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
763                             character > 255 is encountered. */                             character > 255 is encountered. */
764    
765    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
766                             class. This does both positive and negative. */                             class. This does both positive and negative. */
767    
768    OP_REF,            /* 62 Match a back reference */    OP_REF,            /* 81 Match a back reference */
769    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
770    OP_CALLOUT,        /* 64 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
771    
772    OP_ALT,            /* 65 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
773    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 */
774    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
775    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
776    
777    /* The assertions must come before ONCE and COND */    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
778    
779    OP_ASSERT,         /* 69 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
780    OP_ASSERT_NOT,     /* 70 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
781    OP_ASSERTBACK,     /* 71 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
782    OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
783    OP_REVERSE,        /* 73 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
784    
785    /* ONCE and COND must come after the assertions, with ONCE first, as there's    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
786    a test for >= ONCE for a subpattern that isn't an assertion. */    as there's a test for >= ONCE for a subpattern that isn't an assertion. */
787    
788    OP_ONCE,           /* 74 Once matched, don't back up into the subpattern */    OP_ONCE,           /* 93 Atomic group */
789    OP_COND,           /* 75 Conditional group */    OP_BRA,            /* 94 Start of non-capturing bracket */
790    OP_CREF,           /* 76 Used to hold an extraction string number (cond ref) */    OP_CBRA,           /* 95 Start of capturing bracket */
791      OP_COND,           /* 96 Conditional group */
792    OP_BRAZERO,        /* 77 These two must remain together and in this */  
793    OP_BRAMINZERO,     /* 78 order. */    /* These three must follow the previous three, in the same order. There's a
794      check for >= SBRA to distinguish the two sets. */
795    OP_BRANUMBER,      /* 79 Used for extracting brackets whose number is greater  
796                             than can fit into an opcode. */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
797      OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
798    OP_BRA             /* 80 This and greater values are used for brackets that    OP_SCOND,          /* 99 Conditional group, check empty */
                            extract substrings up to EXTRACT_BASIC_MAX. After  
                            that, use is made of OP_BRANUMBER. */  
 };  
799    
800  /* WARNING WARNING WARNING: There is an implicit assumption in pcre.c and    OP_CREF,           /* 100 Used to hold a capture number as condition */
801  study.c that all opcodes are less than 128 in value. This makes handling UTF-8    OP_RREF,           /* 101 Used to hold a recursion number as condition */
802  character sequences easier. */    OP_DEF,            /* 102 The DEFINE condition */
   
 /* 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. */  
803    
804  #define EXTRACT_BASIC_MAX  100    OP_BRAZERO,        /* 103 These two must remain together and in this */
805      OP_BRAMINZERO,     /* 104 order. */
806    
807      /* These are backtracking control verbs */
808    
809      OP_PRUNE,          /* 105 */
810      OP_SKIP,           /* 106 */
811      OP_THEN,           /* 107 */
812      OP_COMMIT,         /* 108 */
813    
814      /* These are forced failure and success verbs */
815    
816      OP_FAIL,           /* 109 */
817      OP_ACCEPT,         /* 110 */
818    
819      /* This is used to skip a subpattern with a {0} quantifier */
820    
821      OP_SKIPZERO        /* 111 */
822    };
823    
824    
825  /* 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
826  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
827    
828  #define OP_NAME_LIST \  #define OP_NAME_LIST \
829    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
830    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
831    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
832    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
833    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
834    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
835      "*+","++", "?+", "{",                                           \
836    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
837      "*+","++", "?+", "{",                                           \
838    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
839      "*+","++", "?+", "{",                                           \
840    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
841    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
842    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
843    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "AssertB", "AssertB not", "Reverse",                            \
844    "Brazero", "Braminzero", "Branumber", "Bra"    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
845      "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
846      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
847      "Skip zero"
848    
849    
850  /* 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 657  in UTF-8 mode. The code that uses this t Line 858  in UTF-8 mode. The code that uses this t
858    
859  #define OP_LENGTHS \  #define OP_LENGTHS \
860    1,                             /* End                                    */ \    1,                             /* End                                    */ \
861    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                     */ \
862    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
863      1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
864    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
865      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
866    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
867    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
868    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 667  in UTF-8 mode. The code that uses this t Line 870  in UTF-8 mode. The code that uses this t
870    /* Positive single-char repeats                            ** These are  */ \    /* Positive single-char repeats                            ** These are  */ \
871    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
872    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
873      2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
874    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
875    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
876    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
877      2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
878    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
879    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
880    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
881      2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
882    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
883    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
884    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
# Line 691  in UTF-8 mode. The code that uses this t Line 897  in UTF-8 mode. The code that uses this t
897    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
898    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
899    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* Reverse                                */ \
900    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
901      1+LINK_SIZE,                   /* BRA                                    */ \
902      3+LINK_SIZE,                   /* CBRA                                   */ \
903    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
904      1+LINK_SIZE,                   /* SBRA                                   */ \
905      3+LINK_SIZE,                   /* SCBRA                                  */ \
906      1+LINK_SIZE,                   /* SCOND                                  */ \
907    3,                             /* CREF                                   */ \    3,                             /* CREF                                   */ \
908      3,                             /* RREF                                   */ \
909      1,                             /* DEF                                    */ \
910    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
911    3,                             /* BRANUMBER                              */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
912    1+LINK_SIZE                    /* BRA                                    */ \    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
913    
914    
915  /* A magic value for OP_CREF to indicate the "in recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
916    
917  #define CREF_RECURSE  0xffff  #define RREF_ANY  0xffff
918    
919  /* 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
920  tracked. */  tracked. */
# Line 710  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 923  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
923         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
924         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
925         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
926         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
927           ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
928           ERR60, ERR61, ERR62, ERR63, ERR64 };
929    
930  /* 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
931  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 732  NOTE NOTE NOTE: Line 947  NOTE NOTE NOTE:
947  typedef struct real_pcre {  typedef struct real_pcre {
948    pcre_uint32 magic_number;    pcre_uint32 magic_number;
949    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
950    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
951    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
952      pcre_uint16 dummy1;             /* For future use */
953    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
954    pcre_uint16 top_backref;    pcre_uint16 top_backref;
955    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 765  typedef struct compile_data { Line 980  typedef struct compile_data {
980    const uschar *fcc;            /* Points to case-flipping table */    const uschar *fcc;            /* Points to case-flipping table */
981    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
982    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
983      const uschar *start_workspace;/* The start of working space */
984    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
985    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
986      const uschar *end_pattern;    /* The end of the pattern */
987      uschar *hwm;                  /* High watermark of workspace */
988    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
989    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
990    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
991      int  bracount;                /* Count of capturing parens as we compile */
992      int  final_bracount;          /* Saved value after first pass */
993    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
994    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
995      int  external_options;        /* External (initial) options */
996      int  external_flags;          /* External flag bits to be set */
997    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
998    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
999      int  nltype;                  /* Newline type */
1000      int  nllen;                   /* Newline string length */
1001      uschar nl[4];                 /* Newline string when fixed length */
1002  } compile_data;  } compile_data;
1003    
1004  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
# Line 791  typedef struct recursion_info { Line 1016  typedef struct recursion_info {
1016    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1017    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1018    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 */
1019    USPTR save_start;             /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
1020    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1021    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1022  } recursion_info;  } recursion_info;
1023    
1024  /* 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
1025  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
1026  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. */
1027  with any changes of local variable. However, the pointer to the current frame  
1028  must be saved in some "static" place over a longjmp(). We declare the  typedef struct eptrblock {
1029  structure here so that we can put a pointer in the match_data structure.    struct eptrblock *epb_prev;
1030  NOTE: This isn't used for a "normal" compilation of pcre. */    USPTR epb_saved_eptr;
1031    } eptrblock;
1032    
 struct heapframe;  
1033    
1034  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1035  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
# Line 816  typedef struct match_data { Line 1041  typedef struct match_data {
1041    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
1042    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
1043    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1044      int    nltype;                /* Newline type */
1045      int    nllen;                 /* Newline string length */
1046      uschar nl[4];                 /* Newline string when fixed */
1047    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1048    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
1049    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
1050    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1051    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1052    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1053      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1054    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1055    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1056    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1057    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1058      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1059    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1060    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1061    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1062    USPTR  start_match;           /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1063    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1064    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1065    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1066    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
1067      eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
1068      int    eptrn;                 /* Next free eptrblock */
1069    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1070    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 */  
1071  } match_data;  } match_data;
1072    
1073  /* 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 849  typedef struct dfa_match_data { Line 1080  typedef struct dfa_match_data {
1080    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1081    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1082    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1083      int    nltype;                /* Newline type */
1084      int    nllen;                 /* Newline string length */
1085      uschar nl[4];                 /* Newline string when fixed */
1086    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1087  } dfa_match_data;  } dfa_match_data;
1088    
# Line 858  typedef struct dfa_match_data { Line 1092  typedef struct dfa_match_data {
1092  #define ctype_letter  0x02  #define ctype_letter  0x02
1093  #define ctype_digit   0x04  #define ctype_digit   0x04
1094  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1095  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1096  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1097    
1098  /* 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 886  total length. */ Line 1120  total length. */
1120  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1121    
1122  /* Layout of the UCP type table that translates property names into types and  /* Layout of the UCP type table that translates property names into types and
1123  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1124    relocations in shared libraries, it now has an offset into a single string
1125    instead. */
1126    
1127  typedef struct {  typedef struct {
1128    const char *name;    pcre_uint16 name_offset;
1129    pcre_uint16 type;    pcre_uint16 type;
1130    pcre_uint16 value;    pcre_uint16 value;
1131  } ucp_type_table;  } ucp_type_table;
# Line 907  extern const uschar _pcre_utf8_table4[]; Line 1143  extern const uschar _pcre_utf8_table4[];
1143    
1144  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1145    
1146    extern const char   _pcre_utt_names[];
1147  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1148  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1149    
# Line 919  extern const uschar _pcre_OP_lengths[]; Line 1156  extern const uschar _pcre_OP_lengths[];
1156  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
1157  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1158    
1159  extern int         _pcre_ord2utf8(int, uschar *);  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
1160  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,                        int *, BOOL);
1161                       const pcre_study_data *, pcre_study_data *);  extern int          _pcre_ord2utf8(int, uschar *);
1162  extern int         _pcre_ucp_findprop(const int, int *, int *);  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1163  extern int         _pcre_ucp_othercase(const int);                        const pcre_study_data *, pcre_study_data *);
1164  extern int         _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1165  extern BOOL        _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1166                          int *, BOOL);
1167    extern BOOL         _pcre_xclass(int, const uschar *);
1168    
1169    
1170    /* Unicode character database (UCD) */
1171    
1172    typedef struct {
1173      uschar script;
1174      uschar chartype;
1175      pcre_int32 other_case;
1176    } ucd_record;
1177    
1178    extern const ucd_record  _pcre_ucd_records[];
1179    extern const uschar      _pcre_ucd_stage1[];
1180    extern const pcre_uint16 _pcre_ucd_stage2[];
1181    extern const int         _pcre_ucp_gentype[];
1182    
1183    
1184    /* UCD access macros */
1185    
1186    #define UCD_BLOCK_SIZE 128
1187    #define GET_UCD(ch) (_pcre_ucd_records + \
1188            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1189            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1190    
1191    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1192    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1193    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1194    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1195    
1196  #endif  #endif
1197    

Legend:
Removed from v.87  
changed lines
  Added in v.389

  ViewVC Help
Powered by ViewVC 1.1.5