/[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 351 by ph10, Fri Jul 4 18:27:16 2008 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) 1997-2008 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  #endif
134    
135  /* 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 140  preprocessor time in standard C environm
140    
141  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
142    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
143      typedef short pcre_int16;
144  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
145    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
146      typedef int pcre_int16;
147  #else  #else
148    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
149  #endif  #endif
150    
151  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
152    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
153      typedef int pcre_int32;
154  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
155    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
156      typedef long int pcre_int32;
157  #else  #else
158    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
159  #endif  #endif
# Line 114  Unix, where it is defined in sys/types, Line 167  Unix, where it is defined in sys/types,
167    
168  typedef unsigned char uschar;  typedef unsigned char uschar;
169    
170    /* This is an unsigned int value that no character can ever have. UTF-8
171    characters only go up to 0x7fffffff (though Unicode doesn't go beyond
172    0x0010ffff). */
173    
174    #define NOTACHAR 0xffffffff
175    
176    /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
177    "any" and "anycrlf" at present). The following macros are used to package up
178    testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
179    modules to indicate in which datablock the parameters exist, and what the
180    start/end of string field names are. */
181    
182    #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
183    #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
184    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
185    
186    /* This macro checks for a newline at the given position */
187    
188    #define IS_NEWLINE(p) \
189      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
190        ((p) < NLBLOCK->PSEND && \
191         _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
192           utf8)) \
193        : \
194        ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
195         (p)[0] == NLBLOCK->nl[0] && \
196         (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
197        ) \
198      )
199    
200    /* This macro checks for a newline immediately preceding the given position */
201    
202    #define WAS_NEWLINE(p) \
203      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
204        ((p) > NLBLOCK->PSSTART && \
205         _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
206           &(NLBLOCK->nllen), utf8)) \
207        : \
208        ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
209         (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
210         (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
211        ) \
212      )
213    
214  /* 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
215  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()
216  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 228  must begin with PCRE_. */
228  #define USPTR const unsigned char *  #define USPTR const unsigned char *
229  #endif  #endif
230    
231    
232    
233  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
234  values. */  values. */
235    
# Line 142  need to have their names changed. PCRE m Line 241  need to have their names changed. PCRE m
241  option on the command line. */  option on the command line. */
242    
243  #ifdef VPCOMPAT  #ifdef VPCOMPAT
244    #define strlen(s)        _strlen(s)
245  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
246    #define memcmp(s,c,n)    _memcmp(s,c,n)
247  #define memcpy(d,s,n)    _memcpy(d,s,n)  #define memcpy(d,s,n)    _memcpy(d,s,n)
248  #define memmove(d,s,n)   _memmove(d,s,n)  #define memmove(d,s,n)   _memmove(d,s,n)
249  #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 252  option on the command line. */
252  /* 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(),
253  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
254  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
255  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. */  
256    
257  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
258  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
259  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
260  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
261  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
262  void *  static void *
263  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
264  {  {
265  size_t i;  size_t i;
266  dest += n;  unsigned char *dest = (unsigned char *)d;
267  src += n;  const unsigned char *src = (const unsigned char *)s;
268  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
269  return dest;    {
270      dest += n;
271      src += n;
272      for (i = 0; i < n; ++i) *(--dest) = *(--src);
273      return (void *)dest;
274      }
275    else
276      {
277      for (i = 0; i < n; ++i) *dest++ = *src++;
278      return (void *)(dest - n);
279      }
280  }  }
281  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
282  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 253  capturing parenthesis numbers in back re Line 362  capturing parenthesis numbers in back re
362    
363  /* 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
364  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
365  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
366    never be called in byte mode. To make sure it can never even appear when UTF-8
367    support is omitted, we don't even define it. */
368    
369  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
370    #define NEXTCHAR(p) p++;
371  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
372  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
373  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
374  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
375  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
376  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
377    
378  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
379    
380    /* Advance a character pointer one byte in non-UTF-8 mode and by one character
381    in UTF-8 mode. */
382    
383    #define NEXTCHAR(p) \
384      p++; \
385      if (utf8) { while((*p & 0xc0) == 0x80) p++; }
386    
387  /* Get the next UTF-8 character, not advancing the pointer. This is called when  /* Get the next UTF-8 character, not advancing the pointer. This is called when
388  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
389    
390  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
391    c = *eptr; \    c = *eptr; \
392    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
393      { \      { \
394      int gcii; \      int gcii; \
395      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 407  pointer. */
407    
408  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
409    c = *eptr; \    c = *eptr; \
410    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
411      { \      { \
412      int gcii; \      int gcii; \
413      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 425  know we are in UTF-8 mode. */
425    
426  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
427    c = *eptr++; \    c = *eptr++; \
428    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
429      { \      { \
430      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
431      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 322  know we are in UTF-8 mode. */ Line 441  know we are in UTF-8 mode. */
441    
442  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
443    c = *eptr++; \    c = *eptr++; \
444    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
445      { \      { \
446      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
447      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 339  if there are extra bytes. This is called Line 458  if there are extra bytes. This is called
458    
459  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
460    c = *eptr; \    c = *eptr; \
461    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
462      { \      { \
463      int gcii; \      int gcii; \
464      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 473  if there are extra bytes. This is called
473      }      }
474    
475  /* 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
476  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
477    because almost all calls are already within a block of UTF-8 only code. */
478    
479  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
480    
481  #endif  #endif
482    
# Line 373  Standard C system should have one. */ Line 493  Standard C system should have one. */
493    
494  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
495    
496  /* Private options flags start at the most significant end of the four bytes,  /* Private flags containing information about the compiled regex. They used to
497  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
498  with negative values. The public options defined in pcre.h start at the least  are in a 16-bit flags word. */
499  significant end. Make sure they don't overlap! */  
500    #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
501  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
502  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
503  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
504  #define PCRE_ICHANGED      0x08000000  /* i option changes within regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
505  #define PCRE_NOPARTIAL     0x04000000  /* can't use partial with this regex */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
506    
507  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
508    
# Line 391  significant end. Make sure they don't ov Line 511  significant end. Make sure they don't ov
511  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
512  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
513    
514    #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
515                               PCRE_NEWLINE_ANYCRLF)
516    
517  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
518    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
519     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
520     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| \
521       PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
522       PCRE_JAVASCRIPT_COMPAT)
523    
524  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
525    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
526     PCRE_PARTIAL)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
527    
528  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
529    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
530     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
531       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
532    
533  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
534    
# Line 434  typedef int BOOL; Line 560  typedef int BOOL;
560  #define FALSE   0  #define FALSE   0
561  #define TRUE    1  #define TRUE    1
562    
563  /* 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). */  
564    
565  #ifndef ESC_e  #ifndef ESC_e
566  #define ESC_e 27  #define ESC_e 27
# Line 447  ESC_n is defined as yet another macro, w Line 571  ESC_n is defined as yet another macro, w
571  #endif  #endif
572    
573  #ifndef ESC_n  #ifndef ESC_n
574  #define ESC_n NEWLINE  #define ESC_n '\n'
575  #endif  #endif
576    
577  #ifndef ESC_r  #ifndef ESC_r
# Line 485  contain UTF-8 characters with values gre Line 609  contain UTF-8 characters with values gre
609  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
610  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
611  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
612  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
613  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 };  
614    
615  /* 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
616  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
617    greater than ESC_b and less than ESC_Z to detect the types that may be
618    repeated. These are the types that consume characters. If any new escapes are
619    put in between that don't consume a character, that code will have to change.
620    */
621    
622    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
623           ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
624           ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
625           ESC_REF };
626    
627    
628    /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
629  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.
630  Note that whenever this list is updated, the two macro definitions that follow  
631  must also be updated to match. */  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
632    that follow must also be updated to match. There is also a table called
633    "coptable" in pcre_dfa_exec.c that must be updated. */
634    
635  enum {  enum {
636    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 509  enum { Line 639  enum {
639    
640    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
641    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
642    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
643    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
644    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
645    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
646    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
647    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
648    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
649    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
650    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
651    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
652    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
653    OP_PROP,           /* 14 \p (Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
654    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
655    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_PROP,           /* 16 \p (Unicode property) */
656    OP_EOD,            /* 17 End of data: \z */    OP_ANYNL,          /* 17 \R (any newline sequence) */
657      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
658    OP_OPT,            /* 18 Set runtime options */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
659    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
660    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
661    OP_CHAR,           /* 21 Match one character, casefully */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
662    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
663    OP_NOT,            /* 23 Match anything but the following char */    OP_EOD,            /* 24 End of data: \z */
664    
665    OP_STAR,           /* 24 The maximizing and minimizing versions of */    OP_OPT,            /* 25 Set runtime options */
666    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
667    OP_PLUS,           /* 26 the minimizing one second. */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
668    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_CHAR,           /* 28 Match one character, casefully */
669    OP_QUERY,          /* 28 */    OP_CHARNC,         /* 29 Match one character, caselessly */
670    OP_MINQUERY,       /* 29 */    OP_NOT,            /* 30 Match one character, not the following one */
671    OP_UPTO,           /* 30 From 0 to n matches */  
672    OP_MINUPTO,        /* 31 */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
673    OP_EXACT,          /* 32 Exactly n matches */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
674      OP_PLUS,           /* 33 the minimizing one second. */
675    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
676    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */    OP_QUERY,          /* 35 */
677    OP_NOTPLUS,        /* 35 the minimizing one second. */    OP_MINQUERY,       /* 36 */
678    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */  
679    OP_NOTQUERY,       /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
680    OP_NOTMINQUERY,    /* 38 */    OP_MINUPTO,        /* 38 */
681    OP_NOTUPTO,        /* 39 From 0 to n matches */    OP_EXACT,          /* 39 Exactly n matches */
682    OP_NOTMINUPTO,     /* 40 */  
683    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_POSSTAR,        /* 40 Possessified star */
684      OP_POSPLUS,        /* 41 Possessified plus */
685    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_POSQUERY,       /* 42 Posesssified query */
686    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */    OP_POSUPTO,        /* 43 Possessified upto */
687    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */  
688    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
689    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
690    OP_TYPEMINQUERY,   /* 47 */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
691    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
692    OP_TYPEMINUPTO,    /* 49 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
693    OP_TYPEEXACT,      /* 50 Exactly n matches */    OP_NOTMINQUERY,    /* 49 */
694    
695    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */    OP_NOTUPTO,        /* 50 From 0 to n matches */
696    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    OP_NOTMINUPTO,     /* 51 */
697    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */    OP_NOTEXACT,       /* 52 Exactly n matches */
698    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */  
699    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
700    OP_CRMINQUERY,     /* 56 */    OP_NOTPOSPLUS,     /* 54 */
701    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_NOTPOSQUERY,    /* 55 */
702    OP_CRMINRANGE,     /* 58 */    OP_NOTPOSUPTO,     /* 56 */
703    
704      OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
705      OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
706      OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
707      OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
708      OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
709      OP_TYPEMINQUERY,   /* 62 */
710    
711      OP_TYPEUPTO,       /* 63 From 0 to n matches */
712      OP_TYPEMINUPTO,    /* 64 */
713      OP_TYPEEXACT,      /* 65 Exactly n matches */
714    
715      OP_TYPEPOSSTAR,    /* 66 Possessified versions */
716      OP_TYPEPOSPLUS,    /* 67 */
717      OP_TYPEPOSQUERY,   /* 68 */
718      OP_TYPEPOSUPTO,    /* 69 */
719    
720      OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
721      OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
722      OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
723      OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
724      OP_CRQUERY,        /* 74 These are for character classes and back refs */
725      OP_CRMINQUERY,     /* 75 */
726      OP_CRRANGE,        /* 76 These are different to the three sets above. */
727      OP_CRMINRANGE,     /* 77 */
728    
729    OP_CLASS,          /* 59 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
730    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
731                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
732                             character > 255 is encountered. */                             character > 255 is encountered. */
733    
734    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
735                             class. This does both positive and negative. */                             class. This does both positive and negative. */
736    
737    OP_REF,            /* 62 Match a back reference */    OP_REF,            /* 81 Match a back reference */
738    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
739    OP_CALLOUT,        /* 64 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
740    
741    OP_ALT,            /* 65 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
742    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 */
743    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
744    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
745    
746    /* The assertions must come before ONCE and COND */    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
747    
748    OP_ASSERT,         /* 69 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
749    OP_ASSERT_NOT,     /* 70 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
750    OP_ASSERTBACK,     /* 71 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
751    OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
752    OP_REVERSE,        /* 73 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
   
   /* ONCE and COND must come after the assertions, with ONCE first, as there's  
   a test for >= ONCE for a subpattern that isn't an assertion. */  
   
   OP_ONCE,           /* 74 Once matched, don't back up into the subpattern */  
   OP_COND,           /* 75 Conditional group */  
   OP_CREF,           /* 76 Used to hold an extraction string number (cond ref) */  
   
   OP_BRAZERO,        /* 77 These two must remain together and in this */  
   OP_BRAMINZERO,     /* 78 order. */  
   
   OP_BRANUMBER,      /* 79 Used for extracting brackets whose number is greater  
                            than can fit into an opcode. */  
   
   OP_BRA             /* 80 This and greater values are used for brackets that  
                            extract substrings up to EXTRACT_BASIC_MAX. After  
                            that, use is made of OP_BRANUMBER. */  
 };  
753    
754  /* WARNING WARNING WARNING: There is an implicit assumption in pcre.c and    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
755  study.c that all opcodes are less than 128 in value. This makes handling UTF-8    as there's a test for >= ONCE for a subpattern that isn't an assertion. */
 character sequences easier. */  
   
 /* The highest extraction number before we have to start using additional  
 bytes. (Originally PCRE didn't have support for extraction counts highter than  
 this number.) The value is limited by the number of opcodes left after OP_BRA,  
 i.e. 255 - OP_BRA. We actually set it a bit lower to leave room for additional  
 opcodes. */  
756    
757  #define EXTRACT_BASIC_MAX  100    OP_ONCE,           /* 93 Atomic group */
758      OP_BRA,            /* 94 Start of non-capturing bracket */
759      OP_CBRA,           /* 95 Start of capturing bracket */
760      OP_COND,           /* 96 Conditional group */
761    
762      /* These three must follow the previous three, in the same order. There's a
763      check for >= SBRA to distinguish the two sets. */
764    
765      OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
766      OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
767      OP_SCOND,          /* 99 Conditional group, check empty */
768    
769      OP_CREF,           /* 100 Used to hold a capture number as condition */
770      OP_RREF,           /* 101 Used to hold a recursion number as condition */
771      OP_DEF,            /* 102 The DEFINE condition */
772    
773      OP_BRAZERO,        /* 103 These two must remain together and in this */
774      OP_BRAMINZERO,     /* 104 order. */
775    
776      /* These are backtracking control verbs */
777    
778      OP_PRUNE,          /* 105 */
779      OP_SKIP,           /* 106 */
780      OP_THEN,           /* 107 */
781      OP_COMMIT,         /* 108 */
782    
783      /* These are forced failure and success verbs */
784    
785      OP_FAIL,           /* 109 */
786      OP_ACCEPT,         /* 110 */
787    
788      /* This is used to skip a subpattern with a {0} quantifier */
789    
790      OP_SKIPZERO        /* 111 */
791    };
792    
793    
794  /* 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
795  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
796    
797  #define OP_NAME_LIST \  #define OP_NAME_LIST \
798    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
799    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
800    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
801    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
802    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
803    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
804      "*+","++", "?+", "{",                                           \
805    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
806      "*+","++", "?+", "{",                                           \
807    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
808      "*+","++", "?+", "{",                                           \
809    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
810    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
811    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
812    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "AssertB", "AssertB not", "Reverse",                            \
813    "Brazero", "Braminzero", "Branumber", "Bra"    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
814      "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
815      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
816      "Skip zero"
817    
818    
819  /* 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 827  in UTF-8 mode. The code that uses this t
827    
828  #define OP_LENGTHS \  #define OP_LENGTHS \
829    1,                             /* End                                    */ \    1,                             /* End                                    */ \
830    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                     */ \
831    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
832      1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
833    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
834      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
835    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
836    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
837    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 667  in UTF-8 mode. The code that uses this t Line 839  in UTF-8 mode. The code that uses this t
839    /* Positive single-char repeats                            ** These are  */ \    /* Positive single-char repeats                            ** These are  */ \
840    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
841    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
842      2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
843    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
844    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
845    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
846      2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
847    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
848    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
849    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
850      2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
851    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
852    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
853    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
# Line 691  in UTF-8 mode. The code that uses this t Line 866  in UTF-8 mode. The code that uses this t
866    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
867    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
868    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* Reverse                                */ \
869    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
870      1+LINK_SIZE,                   /* BRA                                    */ \
871      3+LINK_SIZE,                   /* CBRA                                   */ \
872    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
873      1+LINK_SIZE,                   /* SBRA                                   */ \
874      3+LINK_SIZE,                   /* SCBRA                                  */ \
875      1+LINK_SIZE,                   /* SCOND                                  */ \
876    3,                             /* CREF                                   */ \    3,                             /* CREF                                   */ \
877      3,                             /* RREF                                   */ \
878      1,                             /* DEF                                    */ \
879    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
880    3,                             /* BRANUMBER                              */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
881    1+LINK_SIZE                    /* BRA                                    */ \    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
882    
883    
884  /* A magic value for OP_CREF to indicate the "in recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
885    
886  #define CREF_RECURSE  0xffff  #define RREF_ANY  0xffff
887    
888  /* 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
889  tracked. */  tracked. */
# Line 710  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 892  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
892         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
893         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
894         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
895         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
896           ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
897           ERR60, ERR61, ERR62, ERR63, ERR64 };
898    
899  /* 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
900  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 916  NOTE NOTE NOTE:
916  typedef struct real_pcre {  typedef struct real_pcre {
917    pcre_uint32 magic_number;    pcre_uint32 magic_number;
918    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
919    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
920    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
921      pcre_uint16 dummy1;             /* For future use */
922    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
923    pcre_uint16 top_backref;    pcre_uint16 top_backref;
924    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 765  typedef struct compile_data { Line 949  typedef struct compile_data {
949    const uschar *fcc;            /* Points to case-flipping table */    const uschar *fcc;            /* Points to case-flipping table */
950    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
951    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
952      const uschar *start_workspace;/* The start of working space */
953    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
954    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
955      const uschar *end_pattern;    /* The end of the pattern */
956      uschar *hwm;                  /* High watermark of workspace */
957    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
958    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
959    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
960      int  bracount;                /* Count of capturing parens as we compile */
961      int  final_bracount;          /* Saved value after first pass */
962    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
963    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
964      int  external_options;        /* External (initial) options */
965      int  external_flags;          /* External flag bits to be set */
966    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
967    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
968      int  nltype;                  /* Newline type */
969      int  nllen;                   /* Newline string length */
970      uschar nl[4];                 /* Newline string when fixed length */
971  } compile_data;  } compile_data;
972    
973  /* 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 985  typedef struct recursion_info {
985    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
986    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
987    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 */
988    USPTR save_start;             /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
989    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
990    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
991  } recursion_info;  } recursion_info;
992    
993  /* 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
994  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
995  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. */
996  with any changes of local variable. However, the pointer to the current frame  
997  must be saved in some "static" place over a longjmp(). We declare the  typedef struct eptrblock {
998  structure here so that we can put a pointer in the match_data structure.    struct eptrblock *epb_prev;
999  NOTE: This isn't used for a "normal" compilation of pcre. */    USPTR epb_saved_eptr;
1000    } eptrblock;
1001    
 struct heapframe;  
1002    
1003  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1004  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 1010  typedef struct match_data {
1010    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
1011    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
1012    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1013      int    nltype;                /* Newline type */
1014      int    nllen;                 /* Newline string length */
1015      uschar nl[4];                 /* Newline string when fixed */
1016    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1017    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
1018    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
1019    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1020    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1021    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1022      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1023    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1024    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1025    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1026    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1027      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1028    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1029    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1030    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1031    USPTR  start_match;           /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1032    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1033    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1034    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1035    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
1036      eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
1037      int    eptrn;                 /* Next free eptrblock */
1038    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1039    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 */  
1040  } match_data;  } match_data;
1041    
1042  /* 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 1049  typedef struct dfa_match_data {
1049    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1050    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1051    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1052      int    nltype;                /* Newline type */
1053      int    nllen;                 /* Newline string length */
1054      uschar nl[4];                 /* Newline string when fixed */
1055    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1056  } dfa_match_data;  } dfa_match_data;
1057    
# Line 858  typedef struct dfa_match_data { Line 1061  typedef struct dfa_match_data {
1061  #define ctype_letter  0x02  #define ctype_letter  0x02
1062  #define ctype_digit   0x04  #define ctype_digit   0x04
1063  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1064  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1065  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1066    
1067  /* 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 1089  total length. */
1089  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1090    
1091  /* 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
1092  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1093    relocations in shared libraries, it now has an offset into a single string
1094    instead. */
1095    
1096  typedef struct {  typedef struct {
1097    const char *name;    pcre_uint16 name_offset;
1098    pcre_uint16 type;    pcre_uint16 type;
1099    pcre_uint16 value;    pcre_uint16 value;
1100  } ucp_type_table;  } ucp_type_table;
# Line 907  extern const uschar _pcre_utf8_table4[]; Line 1112  extern const uschar _pcre_utf8_table4[];
1112    
1113  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1114    
1115    extern const char   _pcre_utt_names[];
1116  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1117  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1118    
# Line 919  extern const uschar _pcre_OP_lengths[]; Line 1125  extern const uschar _pcre_OP_lengths[];
1125  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
1126  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1127    
1128  extern int         _pcre_ord2utf8(int, uschar *);  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
1129  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,                        int *, BOOL);
1130                       const pcre_study_data *, pcre_study_data *);  extern int          _pcre_ord2utf8(int, uschar *);
1131  extern int         _pcre_ucp_findprop(const int, int *, int *);  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1132  extern int         _pcre_ucp_othercase(const int);                        const pcre_study_data *, pcre_study_data *);
1133  extern int         _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1134  extern BOOL        _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1135                          int *, BOOL);
1136    extern BOOL         _pcre_xclass(int, const uschar *);
1137    
1138    
1139    /* Unicode character database (UCD) */
1140    
1141    typedef struct {
1142      uschar script;
1143      uschar chartype;
1144      pcre_int32 other_case;
1145    } ucd_record;
1146    
1147    extern const ucd_record  _pcre_ucd_records[];
1148    extern const uschar      _pcre_ucd_stage1[];
1149    extern const pcre_uint16 _pcre_ucd_stage2[];
1150    extern const int         _pcre_ucp_gentype[];
1151    
1152    
1153    /* UCD access macros */
1154    
1155    #define UCD_BLOCK_SIZE 128
1156    #define GET_UCD(ch) (_pcre_ucd_records + \
1157            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1158            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1159    
1160    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1161    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1162    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1163    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1164    
1165  #endif  #endif
1166    

Legend:
Removed from v.87  
changed lines
  Added in v.351

  ViewVC Help
Powered by ViewVC 1.1.5