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

Diff of /code/trunk/pcre_internal.h

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

revision 85 by nigel, Sat Feb 24 21:41:13 2007 UTC revision 336 by ph10, Sat Apr 12 15:59:03 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) 1997-2005 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 114  Unix, where it is defined in sys/types, Line 163  Unix, where it is defined in sys/types,
163    
164  typedef unsigned char uschar;  typedef unsigned char uschar;
165    
166    /* This is an unsigned int value that no character can ever have. UTF-8
167    characters only go up to 0x7fffffff (though Unicode doesn't go beyond
168    0x0010ffff). */
169    
170    #define NOTACHAR 0xffffffff
171    
172    /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
173    "any" and "anycrlf" at present). The following macros are used to package up
174    testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
175    modules to indicate in which datablock the parameters exist, and what the
176    start/end of string field names are. */
177    
178    #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
179    #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
180    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
181    
182    /* This macro checks for a newline at the given position */
183    
184    #define IS_NEWLINE(p) \
185      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
186        ((p) < NLBLOCK->PSEND && \
187         _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
188           utf8)) \
189        : \
190        ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
191         (p)[0] == NLBLOCK->nl[0] && \
192         (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
193        ) \
194      )
195    
196    /* This macro checks for a newline immediately preceding the given position */
197    
198    #define WAS_NEWLINE(p) \
199      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
200        ((p) > NLBLOCK->PSSTART && \
201         _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
202           &(NLBLOCK->nllen), utf8)) \
203        : \
204        ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
205         (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
206         (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
207        ) \
208      )
209    
210    /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
211    with a custom type. This makes it possible, for example, to allow pcre_exec()
212    to process subject strings that are discontinuous by using a smart pointer
213    class. It must always be possible to inspect all of the subject string in
214    pcre_exec() because of the way it backtracks. Two macros are required in the
215    normal case, for sign-unspecified and unsigned char pointers. The former is
216    used for the external interface and appears in pcre.h, which is why its name
217    must begin with PCRE_. */
218    
219    #ifdef CUSTOM_SUBJECT_PTR
220    #define PCRE_SPTR CUSTOM_SUBJECT_PTR
221    #define USPTR CUSTOM_SUBJECT_PTR
222    #else
223    #define PCRE_SPTR const char *
224    #define USPTR const unsigned char *
225    #endif
226    
227    
228    
229  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
230  values. */  values. */
231    
# Line 125  need to have their names changed. PCRE m Line 237  need to have their names changed. PCRE m
237  option on the command line. */  option on the command line. */
238    
239  #ifdef VPCOMPAT  #ifdef VPCOMPAT
240    #define strlen(s)        _strlen(s)
241  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
242    #define memcmp(s,c,n)    _memcmp(s,c,n)
243  #define memcpy(d,s,n)    _memcpy(d,s,n)  #define memcpy(d,s,n)    _memcpy(d,s,n)
244  #define memmove(d,s,n)   _memmove(d,s,n)  #define memmove(d,s,n)   _memmove(d,s,n)
245  #define memset(s,c,n)    _memset(s,c,n)  #define memset(s,c,n)    _memset(s,c,n)
# Line 134  option on the command line. */ Line 248  option on the command line. */
248  /* 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(),
249  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
250  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
251  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. */  
252    
253  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
254  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
255  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
256  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
257  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
258  void *  static void *
259  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
260  {  {
261  int i;  size_t i;
262  dest += n;  unsigned char *dest = (unsigned char *)d;
263  src += n;  const unsigned char *src = (const unsigned char *)s;
264  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
265      {
266      dest += n;
267      src += n;
268      for (i = 0; i < n; ++i) *(--dest) = *(--src);
269      return (void *)dest;
270      }
271    else
272      {
273      for (i = 0; i < n; ++i) *dest++ = *src++;
274      return (void *)(dest - n);
275      }
276  }  }
277  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
278  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 235  capturing parenthesis numbers in back re Line 358  capturing parenthesis numbers in back re
358    
359  /* 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
360  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
361  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
362    never be called in byte mode. To make sure it can never even appear when UTF-8
363    support is omitted, we don't even define it. */
364    
365  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
366    #define NEXTCHAR(p) p++;
367  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
368  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
369  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
370  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
371  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
372  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
373    
374  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
375    
376    /* Advance a character pointer one byte in non-UTF-8 mode and by one character
377    in UTF-8 mode. */
378    
379    #define NEXTCHAR(p) \
380      p++; \
381      if (utf8) { while((*p & 0xc0) == 0x80) p++; }
382    
383  /* 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
384  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
385    
386  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
387    c = *eptr; \    c = *eptr; \
388    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
389      { \      { \
390      int gcii; \      int gcii; \
391      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 270  pointer. */ Line 403  pointer. */
403    
404  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
405    c = *eptr; \    c = *eptr; \
406    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
407      { \      { \
408      int gcii; \      int gcii; \
409      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 288  know we are in UTF-8 mode. */ Line 421  know we are in UTF-8 mode. */
421    
422  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
423    c = *eptr++; \    c = *eptr++; \
424    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
425      { \      { \
426      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
427      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 304  know we are in UTF-8 mode. */ Line 437  know we are in UTF-8 mode. */
437    
438  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
439    c = *eptr++; \    c = *eptr++; \
440    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
441      { \      { \
442      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
443      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 321  if there are extra bytes. This is called Line 454  if there are extra bytes. This is called
454    
455  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
456    c = *eptr; \    c = *eptr; \
457    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
458      { \      { \
459      int gcii; \      int gcii; \
460      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 336  if there are extra bytes. This is called Line 469  if there are extra bytes. This is called
469      }      }
470    
471  /* 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
472  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
473    because almost all calls are already within a block of UTF-8 only code. */
474    
475  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
476    
477  #endif  #endif
478    
# Line 355  Standard C system should have one. */ Line 489  Standard C system should have one. */
489    
490  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
491    
492  /* Private options flags start at the most significant end of the four bytes,  /* Private flags containing information about the compiled regex. They used to
493  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
494  with negative values. The public options defined in pcre.h start at the least  are in a 16-bit flags word. */
495  significant end. Make sure they don't overlap! */  
496    #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
497  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
498  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
499  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
500  #define PCRE_ICHANGED      0x08000000  /* i option changes within regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
501  #define PCRE_NOPARTIAL     0x04000000  /* can't use partial with this regex */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
502    
503  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
504    
# Line 373  significant end. Make sure they don't ov Line 507  significant end. Make sure they don't ov
507  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
508  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
509    
510    #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
511                               PCRE_NEWLINE_ANYCRLF)
512    
513  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
514    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
515     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
516     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| \
517       PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
518       PCRE_JAVASCRIPT_COMPAT)
519    
520  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
521    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
522     PCRE_PARTIAL)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
523    
524  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_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_DFA_SHORTEST|PCRE_DFA_RESTART)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
527       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
528    
529  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
530    
# Line 416  typedef int BOOL; Line 556  typedef int BOOL;
556  #define FALSE   0  #define FALSE   0
557  #define TRUE    1  #define TRUE    1
558    
559  /* 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). */  
560    
561  #ifndef ESC_e  #ifndef ESC_e
562  #define ESC_e 27  #define ESC_e 27
# Line 429  ESC_n is defined as yet another macro, w Line 567  ESC_n is defined as yet another macro, w
567  #endif  #endif
568    
569  #ifndef ESC_n  #ifndef ESC_n
570  #define ESC_n NEWLINE  #define ESC_n '\n'
571  #endif  #endif
572    
573  #ifndef ESC_r  #ifndef ESC_r
# Line 443  ESC_n is defined as yet another macro, w Line 581  ESC_n is defined as yet another macro, w
581  #define ESC_tee '\t'  #define ESC_tee '\t'
582  #endif  #endif
583    
584  /* These are escaped items that aren't just an encoding of a particular data  /* Codes for different types of Unicode property */
 value such as \n. They must have non-zero values, as check_escape() returns  
 their negation. Also, they must appear in the same order as in the opcode  
 definitions below, up to ESC_z. There's a dummy for OP_ANY because it  
 corresponds to "." rather than an escape sequence. The final one must be  
 ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two  
 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. */  
585    
586  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  #define PT_ANY        0    /* Any property - matches all chars */
587         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E,  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
588         ESC_Q, ESC_REF };  #define PT_GC         2    /* General characteristic (e.g. L) */
589    #define PT_PC         3    /* Particular characteristic (e.g. Lu) */
590    #define PT_SC         4    /* Script (e.g. Han) */
591    
592  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
593  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 467  contain UTF-8 characters with values gre Line 598  contain UTF-8 characters with values gre
598  #define XCL_END       0    /* Marks end of individual items */  #define XCL_END       0    /* Marks end of individual items */
599  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */
600  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */
601  #define XCL_PROP      3    /* Unicode property (one property code) follows */  #define XCL_PROP      3    /* Unicode property (2-byte property code follows) */
602  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
603    
604    /* These are escaped items that aren't just an encoding of a particular data
605    value such as \n. They must have non-zero values, as check_escape() returns
606    their negation. Also, they must appear in the same order as in the opcode
607    definitions below, up to ESC_z. There's a dummy for OP_ANY because it
608    corresponds to "." rather than an escape sequence. The final one must be
609    ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).
610    There are two tests in the code for an escape greater than ESC_b and less than
611    ESC_Z to detect the types that may be repeated. These are the types that
612    consume characters. If any new escapes are put in between that don't consume a
613    character, that code will have to change. */
614    
615    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
616           ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,
617           ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k, ESC_REF };
618    
619  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  
620  that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
621  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.
622  Note that whenever this list is updated, the two macro definitions that follow  
623  must also be updated to match. */  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
624    that follow must also be updated to match. There is also a table called
625    "coptable" in pcre_dfa_exec.c that must be updated. */
626    
627  enum {  enum {
628    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 484  enum { Line 631  enum {
631    
632    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
633    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
634    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
635    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
636    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
637    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
638    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
639    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
640    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
641    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
642    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
643    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character */
644    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */
645    OP_PROP,           /* 14 \p (Unicode property) */    OP_NOTPROP,        /* 14 \P (not Unicode property) */
646    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_PROP,           /* 15 \p (Unicode property) */
647    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_ANYNL,          /* 16 \R (any newline sequence) */
648    OP_EOD,            /* 17 End of data: \z */    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */
649      OP_HSPACE,         /* 18 \h (horizontal whitespace) */
650    OP_OPT,            /* 18 Set runtime options */    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */
651    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_VSPACE,         /* 20 \v (vertical whitespace) */
652    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */
653    OP_CHAR,           /* 21 Match one character, casefully */    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */
654    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_EOD,            /* 23 End of data: \z */
655    OP_NOT,            /* 23 Match anything but the following char */  
656      OP_OPT,            /* 24 Set runtime options */
657    OP_STAR,           /* 24 The maximizing and minimizing versions of */    OP_CIRC,           /* 25 Start of line - varies with multiline switch */
658    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_DOLL,           /* 26 End of line - varies with multiline switch */
659    OP_PLUS,           /* 26 the minimizing one second. */    OP_CHAR,           /* 27 Match one character, casefully */
660    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_CHARNC,         /* 28 Match one character, caselessly */
661    OP_QUERY,          /* 28 */    OP_NOT,            /* 29 Match one character, not the following one */
662    OP_MINQUERY,       /* 29 */  
663    OP_UPTO,           /* 30 From 0 to n matches */    OP_STAR,           /* 30 The maximizing and minimizing versions of */
664    OP_MINUPTO,        /* 31 */    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */
665    OP_EXACT,          /* 32 Exactly n matches */    OP_PLUS,           /* 32 the minimizing one second. */
666      OP_MINPLUS,        /* 33 This first set applies to single characters.*/
667    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    OP_QUERY,          /* 34 */
668    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */    OP_MINQUERY,       /* 35 */
669    OP_NOTPLUS,        /* 35 the minimizing one second. */  
670    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */    OP_UPTO,           /* 36 From 0 to n matches */
671    OP_NOTQUERY,       /* 37 */    OP_MINUPTO,        /* 37 */
672    OP_NOTMINQUERY,    /* 38 */    OP_EXACT,          /* 38 Exactly n matches */
673    OP_NOTUPTO,        /* 39 From 0 to n matches */  
674    OP_NOTMINUPTO,     /* 40 */    OP_POSSTAR,        /* 39 Possessified star */
675    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_POSPLUS,        /* 40 Possessified plus */
676      OP_POSQUERY,       /* 41 Posesssified query */
677    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_POSUPTO,        /* 42 Possessified upto */
678    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */  
679    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */
680    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */
681    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */
682    OP_TYPEMINQUERY,   /* 47 */    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */
683    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */
684    OP_TYPEMINUPTO,    /* 49 */    OP_NOTMINQUERY,    /* 48 */
685    OP_TYPEEXACT,      /* 50 Exactly n matches */  
686      OP_NOTUPTO,        /* 49 From 0 to n matches */
687    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */    OP_NOTMINUPTO,     /* 50 */
688    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    OP_NOTEXACT,       /* 51 Exactly n matches */
689    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */  
690    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */    OP_NOTPOSSTAR,     /* 52 Possessified versions */
691    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_NOTPOSPLUS,     /* 53 */
692    OP_CRMINQUERY,     /* 56 */    OP_NOTPOSQUERY,    /* 54 */
693    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_NOTPOSUPTO,     /* 55 */
694    OP_CRMINRANGE,     /* 58 */  
695      OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */
696      OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */
697      OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */
698      OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */
699      OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */
700      OP_TYPEMINQUERY,   /* 61 */
701    
702      OP_TYPEUPTO,       /* 62 From 0 to n matches */
703      OP_TYPEMINUPTO,    /* 63 */
704      OP_TYPEEXACT,      /* 64 Exactly n matches */
705    
706      OP_TYPEPOSSTAR,    /* 65 Possessified versions */
707      OP_TYPEPOSPLUS,    /* 66 */
708      OP_TYPEPOSQUERY,   /* 67 */
709      OP_TYPEPOSUPTO,    /* 68 */
710    
711      OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */
712      OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */
713      OP_CRPLUS,         /* 71 the minimizing one second. These codes must */
714      OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */
715      OP_CRQUERY,        /* 73 These are for character classes and back refs */
716      OP_CRMINQUERY,     /* 74 */
717      OP_CRRANGE,        /* 75 These are different to the three sets above. */
718      OP_CRMINRANGE,     /* 76 */
719    
720    OP_CLASS,          /* 59 Match a character class, chars < 256 only */    OP_CLASS,          /* 77 Match a character class, chars < 256 only */
721    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative
722                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
723                             character > 255 is encountered. */                             character > 255 is encountered. */
724    
725    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the
726                             class. This does both positive and negative. */                             class. This does both positive and negative. */
727    
728    OP_REF,            /* 62 Match a back reference */    OP_REF,            /* 80 Match a back reference */
729    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */
730    OP_CALLOUT,        /* 64 Call out to external function if provided */    OP_CALLOUT,        /* 82 Call out to external function if provided */
731    
732    OP_ALT,            /* 65 Start of alternation */    OP_ALT,            /* 83 Start of alternation */
733    OP_KET,            /* 66 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */
734    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_KETRMAX,        /* 85 These two must remain together and in this */
735    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */
736    
737    /* The assertions must come before ONCE and COND */    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
738    
739    OP_ASSERT,         /* 69 Positive lookahead */    OP_ASSERT,         /* 87 Positive lookahead */
740    OP_ASSERT_NOT,     /* 70 Negative lookahead */    OP_ASSERT_NOT,     /* 88 Negative lookahead */
741    OP_ASSERTBACK,     /* 71 Positive lookbehind */    OP_ASSERTBACK,     /* 89 Positive lookbehind */
742    OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */
743    OP_REVERSE,        /* 73 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */
744    
745    /* 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,
746    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. */
747    
748    OP_ONCE,           /* 74 Once matched, don't back up into the subpattern */    OP_ONCE,           /* 92 Atomic group */
749    OP_COND,           /* 75 Conditional group */    OP_BRA,            /* 93 Start of non-capturing bracket */
750    OP_CREF,           /* 76 Used to hold an extraction string number (cond ref) */    OP_CBRA,           /* 94 Start of capturing bracket */
751      OP_COND,           /* 95 Conditional group */
752    OP_BRAZERO,        /* 77 These two must remain together and in this */  
753    OP_BRAMINZERO,     /* 78 order. */    /* These three must follow the previous three, in the same order. There's a
754      check for >= SBRA to distinguish the two sets. */
755    OP_BRANUMBER,      /* 79 Used for extracting brackets whose number is greater  
756                             than can fit into an opcode. */    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */
757      OP_SCBRA,          /* 97 Start of capturing bracket, check empty */
758    OP_BRA             /* 80 This and greater values are used for brackets that    OP_SCOND,          /* 98 Conditional group, check empty */
759                             extract substrings up to EXTRACT_BASIC_MAX. After  
760                             that, use is made of OP_BRANUMBER. */    OP_CREF,           /* 99 Used to hold a capture number as condition */
761      OP_RREF,           /* 100 Used to hold a recursion number as condition */
762      OP_DEF,            /* 101 The DEFINE condition */
763    
764      OP_BRAZERO,        /* 102 These two must remain together and in this */
765      OP_BRAMINZERO,     /* 103 order. */
766    
767      /* These are backtracking control verbs */
768    
769      OP_PRUNE,          /* 104 */
770      OP_SKIP,           /* 105 */
771      OP_THEN,           /* 106 */
772      OP_COMMIT,         /* 107 */
773    
774      /* These are forced failure and success verbs */
775    
776      OP_FAIL,           /* 108 */
777      OP_ACCEPT,         /* 109 */
778    
779      /* This is used to skip a subpattern with a {0} quantifier */
780    
781      OP_SKIPZERO        /* 110 */
782  };  };
783    
 /* WARNING WARNING WARNING: There is an implicit assumption in pcre.c and  
 study.c that all opcodes are less than 128 in value. This makes handling UTF-8  
 character sequences easier. */  
   
 /* The highest extraction number before we have to start using additional  
 bytes. (Originally PCRE didn't have support for extraction counts highter than  
 this number.) The value is limited by the number of opcodes left after OP_BRA,  
 i.e. 255 - OP_BRA. We actually set it a bit lower to leave room for additional  
 opcodes. */  
   
 #define EXTRACT_BASIC_MAX  100  
   
784    
785  /* 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
786  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
787    
788  #define OP_NAME_LIST \  #define OP_NAME_LIST \
789    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
790    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \
791    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
792    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
793    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
794    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
795      "*+","++", "?+", "{",                                           \
796    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
797      "*+","++", "?+", "{",                                           \
798    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
799      "*+","++", "?+", "{",                                           \
800    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
801    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
802    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
803    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "AssertB", "AssertB not", "Reverse",                            \
804    "Brazero", "Braminzero", "Branumber", "Bra"    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
805      "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
806      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
807      "Skip zero"
808    
809    
810  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 632  in UTF-8 mode. The code that uses this t Line 818  in UTF-8 mode. The code that uses this t
818    
819  #define OP_LENGTHS \  #define OP_LENGTHS \
820    1,                             /* End                                    */ \    1,                             /* End                                    */ \
821    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                     */ \
822      1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
823    1, 1,                          /* Any, Anybyte                           */ \    1, 1,                          /* Any, Anybyte                           */ \
824    2, 2, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
825      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
826    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
827    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
828    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 642  in UTF-8 mode. The code that uses this t Line 830  in UTF-8 mode. The code that uses this t
830    /* Positive single-char repeats                            ** These are  */ \    /* Positive single-char repeats                            ** These are  */ \
831    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
832    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
833      2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
834    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
835    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
836    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
837      2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
838    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
839    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
840    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
841      2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
842    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
843    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
844    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
# Line 666  in UTF-8 mode. The code that uses this t Line 857  in UTF-8 mode. The code that uses this t
857    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
858    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
859    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* Reverse                                */ \
860    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
861      1+LINK_SIZE,                   /* BRA                                    */ \
862      3+LINK_SIZE,                   /* CBRA                                   */ \
863    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
864      1+LINK_SIZE,                   /* SBRA                                   */ \
865      3+LINK_SIZE,                   /* SCBRA                                  */ \
866      1+LINK_SIZE,                   /* SCOND                                  */ \
867    3,                             /* CREF                                   */ \    3,                             /* CREF                                   */ \
868      3,                             /* RREF                                   */ \
869      1,                             /* DEF                                    */ \
870    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
871    3,                             /* BRANUMBER                              */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
872    1+LINK_SIZE                    /* BRA                                    */ \    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
873    
874    
875  /* A magic value for OP_CREF to indicate the "in recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
876    
877  #define CREF_RECURSE  0xffff  #define RREF_ANY  0xffff
878    
879  /* 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
880  tracked. */  tracked. */
# Line 685  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 883  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
883         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
884         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
885         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
886         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
887           ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
888           ERR60, ERR61, ERR62, ERR63, ERR64 };
889    
890  /* 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
891  code vector run on as long as necessary after the end. We store an explicit  code vector run on as long as necessary after the end. We store an explicit
# Line 707  NOTE NOTE NOTE: Line 907  NOTE NOTE NOTE:
907  typedef struct real_pcre {  typedef struct real_pcre {
908    pcre_uint32 magic_number;    pcre_uint32 magic_number;
909    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
910    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
911    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
912      pcre_uint16 dummy1;             /* For future use */
913    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
914    pcre_uint16 top_backref;    pcre_uint16 top_backref;
915    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 740  typedef struct compile_data { Line 940  typedef struct compile_data {
940    const uschar *fcc;            /* Points to case-flipping table */    const uschar *fcc;            /* Points to case-flipping table */
941    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
942    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
943      const uschar *start_workspace;/* The start of working space */
944    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
945    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
946      const uschar *end_pattern;    /* The end of the pattern */
947      uschar *hwm;                  /* High watermark of workspace */
948    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
949    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
950    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
951      int  bracount;                /* Count of capturing parens as we compile */
952      int  final_bracount;          /* Saved value after first pass */
953    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
954    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
955      int  external_options;        /* External (initial) options */
956      int  external_flags;          /* External flag bits to be set */
957    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
958    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
959      int  nltype;                  /* Newline type */
960      int  nllen;                   /* Newline string length */
961      uschar nl[4];                 /* Newline string when fixed length */
962  } compile_data;  } compile_data;
963    
964  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
# Line 766  typedef struct recursion_info { Line 976  typedef struct recursion_info {
976    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
977    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
978    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 */
979    const uschar *save_start;     /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
980    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
981    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
982  } recursion_info;  } recursion_info;
983    
984  /* 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
985  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
986  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. */
987  with any changes of local variable. However, the pointer to the current frame  
988  must be saved in some "static" place over a longjmp(). We declare the  typedef struct eptrblock {
989  structure here so that we can put a pointer in the match_data structure.    struct eptrblock *epb_prev;
990  NOTE: This isn't used for a "normal" compilation of pcre. */    USPTR epb_saved_eptr;
991    } eptrblock;
992    
 struct heapframe;  
993    
994  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
995  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
996    
997  typedef struct match_data {  typedef struct match_data {
998    unsigned long int match_call_count; /* As it says */    unsigned long int match_call_count;      /* As it says */
999    unsigned long int match_limit;/* As it says */    unsigned long int match_limit;           /* As it says */
1000      unsigned long int match_limit_recursion; /* As it says */
1001    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
1002    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
1003    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1004      int    nltype;                /* Newline type */
1005      int    nllen;                 /* Newline string length */
1006      uschar nl[4];                 /* Newline string when fixed */
1007    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1008    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
1009    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
1010    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1011    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1012    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1013      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1014    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1015    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1016    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1017    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1018      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1019    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1020    const uschar *start_subject;  /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1021    const uschar *end_subject;    /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1022    const uschar *start_match;    /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1023    const uschar *end_match_ptr;  /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1024    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1025    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1026    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
1027      eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
1028      int    eptrn;                 /* Next free eptrblock */
1029    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1030    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 */  
1031  } match_data;  } match_data;
1032    
1033  /* A similar structure is used for the same purpose by the DFA matching  /* A similar structure is used for the same purpose by the DFA matching
# Line 823  typedef struct dfa_match_data { Line 1040  typedef struct dfa_match_data {
1040    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1041    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1042    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1043      int    nltype;                /* Newline type */
1044      int    nllen;                 /* Newline string length */
1045      uschar nl[4];                 /* Newline string when fixed */
1046    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1047  } dfa_match_data;  } dfa_match_data;
1048    
# Line 832  typedef struct dfa_match_data { Line 1052  typedef struct dfa_match_data {
1052  #define ctype_letter  0x02  #define ctype_letter  0x02
1053  #define ctype_digit   0x04  #define ctype_digit   0x04
1054  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1055  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1056  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1057    
1058  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
# Line 859  total length. */ Line 1079  total length. */
1079  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
1080  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1081    
1082  /* Layout of the UCP type table that translates property names into codes for  /* Layout of the UCP type table that translates property names into types and
1083  pcre_ucp_findchar(). */  codes. Each entry used to point directly to a name, but to reduce the number of
1084    relocations in shared libraries, it now has an offset into a single string
1085    instead. */
1086    
1087  typedef struct {  typedef struct {
1088    const char *name;    pcre_uint16 name_offset;
1089    int value;    pcre_uint16 type;
1090      pcre_uint16 value;
1091  } ucp_type_table;  } ucp_type_table;
1092    
1093    
# Line 880  extern const uschar _pcre_utf8_table4[]; Line 1103  extern const uschar _pcre_utf8_table4[];
1103    
1104  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1105    
1106    extern const char   _pcre_utt_names[];
1107  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1108  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1109    
# Line 892  extern const uschar _pcre_OP_lengths[]; Line 1116  extern const uschar _pcre_OP_lengths[];
1116  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
1117  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1118    
1119  extern int         _pcre_ord2utf8(int, uschar *);  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
1120  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,                        int *, BOOL);
1121                       const pcre_study_data *, pcre_study_data *);  extern int          _pcre_ord2utf8(int, uschar *);
1122  extern int         _pcre_ucp_findchar(const int, int *, int *);  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1123  extern int         _pcre_valid_utf8(const uschar *, int);                        const pcre_study_data *, pcre_study_data *);
1124  extern BOOL        _pcre_xclass(int, const uschar *);  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);
1125    extern unsigned int _pcre_ucp_othercase(const unsigned int);
1126    extern int          _pcre_valid_utf8(const uschar *, int);
1127    extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1128                          int *, BOOL);
1129    extern BOOL         _pcre_xclass(int, const uschar *);
1130    
1131  #endif  #endif
1132    

Legend:
Removed from v.85  
changed lines
  Added in v.336

  ViewVC Help
Powered by ViewVC 1.1.5