/[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 118 by ph10, Fri Mar 9 16:17:46 2007 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) 1997-2007 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    
# Line 114  Unix, where it is defined in sys/types, Line 118  Unix, where it is defined in sys/types,
118    
119  typedef unsigned char uschar;  typedef unsigned char uschar;
120    
121    /* This is an unsigned int value that no character can ever have. UTF-8
122    characters only go up to 0x7fffffff (though Unicode doesn't go beyond
123    0x0010ffff). */
124    
125    #define NOTACHAR 0xffffffff
126    
127    /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
128    and "all" at present). The following macros are used to package up testing for
129    newlines. NLBLOCK, PSSTART, and PSEND are defined in the various modules to
130    indicate in which datablock the parameters exist, and what the start/end of
131    string field names are. */
132    
133    #define NLTYPE_FIXED   0     /* Newline is a fixed length string */
134    #define NLTYPE_ANY     1     /* Newline is any Unicode line ending */
135    
136    /* This macro checks for a newline at the given position */
137    
138    #define IS_NEWLINE(p) \
139      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
140        ((p) < NLBLOCK->PSEND && \
141         _pcre_is_newline((p), NLBLOCK->PSEND, &(NLBLOCK->nllen), utf8) \
142        ) \
143        : \
144        ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
145         (p)[0] == NLBLOCK->nl[0] && \
146         (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
147        ) \
148      )
149    
150    /* This macro checks for a newline immediately preceding the given position */
151    
152    #define WAS_NEWLINE(p) \
153      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
154        ((p) > NLBLOCK->PSSTART && \
155         _pcre_was_newline((p), NLBLOCK->PSSTART, &(NLBLOCK->nllen), utf8) \
156        ) \
157        : \
158        ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
159         (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
160         (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
161        ) \
162      )
163    
164    /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
165    with a custom type. This makes it possible, for example, to allow pcre_exec()
166    to process subject strings that are discontinuous by using a smart pointer
167    class. It must always be possible to inspect all of the subject string in
168    pcre_exec() because of the way it backtracks. Two macros are required in the
169    normal case, for sign-unspecified and unsigned char pointers. The former is
170    used for the external interface and appears in pcre.h, which is why its name
171    must begin with PCRE_. */
172    
173    #ifdef CUSTOM_SUBJECT_PTR
174    #define PCRE_SPTR CUSTOM_SUBJECT_PTR
175    #define USPTR CUSTOM_SUBJECT_PTR
176    #else
177    #define PCRE_SPTR const char *
178    #define USPTR const unsigned char *
179    #endif
180    
181  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
182  values. */  values. */
183    
# Line 125  need to have their names changed. PCRE m Line 189  need to have their names changed. PCRE m
189  option on the command line. */  option on the command line. */
190    
191  #ifdef VPCOMPAT  #ifdef VPCOMPAT
192    #define strlen(s)        _strlen(s)
193  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
194    #define memcmp(s,c,n)    _memcmp(s,c,n)
195  #define memcpy(d,s,n)    _memcpy(d,s,n)  #define memcpy(d,s,n)    _memcpy(d,s,n)
196  #define memmove(d,s,n)   _memmove(d,s,n)  #define memmove(d,s,n)   _memmove(d,s,n)
197  #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 200  option on the command line. */
200  /* 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(),
201  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
202  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
203  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. */  
204    
205  #if ! HAVE_MEMMOVE  #if ! HAVE_MEMMOVE
206  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
207  #if HAVE_BCOPY  #if HAVE_BCOPY
208  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
209  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
210  void *  static void *
211  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
212  {  {
213  int i;  size_t i;
214  dest += n;  unsigned char *dest = (unsigned char *)d;
215  src += n;  const unsigned char *src = (const unsigned char *)s;
216  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
217      {
218      dest += n;
219      src += n;
220      for (i = 0; i < n; ++i) *(--dest) = *(--src);
221      return (void *)dest;
222      }
223    else
224      {
225      for (i = 0; i < n; ++i) *dest++ = *src++;
226      return (void *)(dest - n);
227      }
228  }  }
229  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
230  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 252  we know we are in UTF-8 mode. */ Line 327  we know we are in UTF-8 mode. */
327    
328  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
329    c = *eptr; \    c = *eptr; \
330    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
331      { \      { \
332      int gcii; \      int gcii; \
333      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 345  pointer. */
345    
346  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
347    c = *eptr; \    c = *eptr; \
348    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
349      { \      { \
350      int gcii; \      int gcii; \
351      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 363  know we are in UTF-8 mode. */
363    
364  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
365    c = *eptr++; \    c = *eptr++; \
366    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
367      { \      { \
368      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
369      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 304  know we are in UTF-8 mode. */ Line 379  know we are in UTF-8 mode. */
379    
380  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
381    c = *eptr++; \    c = *eptr++; \
382    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
383      { \      { \
384      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
385      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 321  if there are extra bytes. This is called Line 396  if there are extra bytes. This is called
396    
397  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
398    c = *eptr; \    c = *eptr; \
399    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
400      { \      { \
401      int gcii; \      int gcii; \
402      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 355  Standard C system should have one. */ Line 430  Standard C system should have one. */
430    
431  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
432    
433  /* Private options flags start at the most significant end of the four bytes,  /* Private options flags start at the most significant end of the four bytes.
434  but skip the top bit so we can use ints for convenience without getting tangled  The public options defined in pcre.h start at the least significant end. Make
435  with negative values. The public options defined in pcre.h start at the least  sure they don't overlap! The bits are getting a bit scarce now -- when we run
436  significant end. Make sure they don't overlap! */  out, there is a dummy word in the structure that could be used for the private
437    bits. */
438    
439    #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */
440  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */
441  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */
442  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */
443  #define PCRE_ICHANGED      0x08000000  /* i option changes within regex */  #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */
 #define PCRE_NOPARTIAL     0x04000000  /* can't use partial with this regex */  
444    
445  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
446    
# Line 373  significant end. Make sure they don't ov Line 449  significant end. Make sure they don't ov
449  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
450  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
451    
452    #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY)
453    
454  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
455    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
456     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
457     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| \
458       PCRE_DUPNAMES|PCRE_NEWLINE_BITS)
459    
460  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
461    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
462     PCRE_PARTIAL)     PCRE_PARTIAL|PCRE_NEWLINE_BITS)
463    
464  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
465    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
466     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)
467    
468  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
469    
# Line 416  typedef int BOOL; Line 495  typedef int BOOL;
495  #define FALSE   0  #define FALSE   0
496  #define TRUE    1  #define TRUE    1
497    
498  /* 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). */  
499    
500  #ifndef ESC_e  #ifndef ESC_e
501  #define ESC_e 27  #define ESC_e 27
# Line 429  ESC_n is defined as yet another macro, w Line 506  ESC_n is defined as yet another macro, w
506  #endif  #endif
507    
508  #ifndef ESC_n  #ifndef ESC_n
509  #define ESC_n NEWLINE  #define ESC_n '\n'
510  #endif  #endif
511    
512  #ifndef ESC_r  #ifndef ESC_r
# Line 443  ESC_n is defined as yet another macro, w Line 520  ESC_n is defined as yet another macro, w
520  #define ESC_tee '\t'  #define ESC_tee '\t'
521  #endif  #endif
522    
523  /* 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. */  
524    
525  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 */
526         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 */
527         ESC_Q, ESC_REF };  #define PT_GC         2    /* General characteristic (e.g. L) */
528    #define PT_PC         3    /* Particular characteristic (e.g. Lu) */
529    #define PT_SC         4    /* Script (e.g. Han) */
530    
531  /* 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
532  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 537  contain UTF-8 characters with values gre
537  #define XCL_END       0    /* Marks end of individual items */  #define XCL_END       0    /* Marks end of individual items */
538  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */
539  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */
540  #define XCL_PROP      3    /* Unicode property (one property code) follows */  #define XCL_PROP      3    /* Unicode property (2-byte property code follows) */
541  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
542    
543    /* These are escaped items that aren't just an encoding of a particular data
544    value such as \n. They must have non-zero values, as check_escape() returns
545    their negation. Also, they must appear in the same order as in the opcode
546    definitions below, up to ESC_z. There's a dummy for OP_ANY because it
547    corresponds to "." rather than an escape sequence. The final one must be
548    ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).
549    There are two tests in the code for an escape greater than ESC_b and less than
550    ESC_Z to detect the types that may be repeated. These are the types that
551    consume characters. If any new escapes are put in between that don't consume a
552    character, that code will have to change. */
553    
554    enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,
555           ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,
556           ESC_E, ESC_Q, ESC_k, ESC_REF };
557    
558    
559  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
560  that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
561  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.
562  Note that whenever this list is updated, the two macro definitions that follow  
563  must also be updated to match. */  To keep stored, compiled patterns compatible, new opcodes should be added
564    immediately before OP_BRA, where (since release 7.0) a gap is left for this
565    purpose.
566    
567    *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
568    that follow must also be updated to match. There is also a table called
569    "coptable" in pcre_dfa_exec.c that must be updated. */
570    
571  enum {  enum {
572    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 496  enum { Line 587  enum {
587    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */
588    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_NOTPROP,        /* 13 \P (not Unicode property) */
589    OP_PROP,           /* 14 \p (Unicode property) */    OP_PROP,           /* 14 \p (Unicode property) */
590    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_ANYNL,          /* 15 \R (any newline sequence) */
591    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */
592    OP_EOD,            /* 17 End of data: \z */    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */
593      OP_EOD,            /* 18 End of data: \z */
594    OP_OPT,            /* 18 Set runtime options */  
595    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_OPT,            /* 19 Set runtime options */
596    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_CIRC,           /* 20 Start of line - varies with multiline switch */
597    OP_CHAR,           /* 21 Match one character, casefully */    OP_DOLL,           /* 21 End of line - varies with multiline switch */
598    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_CHAR,           /* 22 Match one character, casefully */
599    OP_NOT,            /* 23 Match anything but the following char */    OP_CHARNC,         /* 23 Match one character, caselessly */
600      OP_NOT,            /* 24 Match one character, not the following one */
601    OP_STAR,           /* 24 The maximizing and minimizing versions of */  
602    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_STAR,           /* 25 The maximizing and minimizing versions of */
603    OP_PLUS,           /* 26 the minimizing one second. */    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */
604    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_PLUS,           /* 27 the minimizing one second. */
605    OP_QUERY,          /* 28 */    OP_MINPLUS,        /* 28 This first set applies to single characters.*/
606    OP_MINQUERY,       /* 29 */    OP_QUERY,          /* 29 */
607    OP_UPTO,           /* 30 From 0 to n matches */    OP_MINQUERY,       /* 30 */
608    OP_MINUPTO,        /* 31 */  
609    OP_EXACT,          /* 32 Exactly n matches */    OP_UPTO,           /* 31 From 0 to n matches */
610      OP_MINUPTO,        /* 32 */
611    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    OP_EXACT,          /* 33 Exactly n matches */
612    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */  
613    OP_NOTPLUS,        /* 35 the minimizing one second. */    OP_POSSTAR,        /* 34 Possessified star */
614    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */    OP_POSPLUS,        /* 35 Possessified plus */
615    OP_NOTQUERY,       /* 37 */    OP_POSQUERY,       /* 36 Posesssified query */
616    OP_NOTMINQUERY,    /* 38 */    OP_POSUPTO,        /* 37 Possessified upto */
617    OP_NOTUPTO,        /* 39 From 0 to n matches */  
618    OP_NOTMINUPTO,     /* 40 */    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */
619    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */
620      OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */
621    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */
622    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */
623    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */    OP_NOTMINQUERY,    /* 43 */
624    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */  
625    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_NOTUPTO,        /* 44 From 0 to n matches */
626    OP_TYPEMINQUERY,   /* 47 */    OP_NOTMINUPTO,     /* 45 */
627    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_NOTEXACT,       /* 46 Exactly n matches */
628    OP_TYPEMINUPTO,    /* 49 */  
629    OP_TYPEEXACT,      /* 50 Exactly n matches */    OP_NOTPOSSTAR,     /* 47 Possessified versions */
630      OP_NOTPOSPLUS,     /* 48 */
631    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */    OP_NOTPOSQUERY,    /* 49 */
632    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    OP_NOTPOSUPTO,     /* 50 */
633    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */  
634    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */
635    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */
636    OP_CRMINQUERY,     /* 56 */    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */
637    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */
638    OP_CRMINRANGE,     /* 58 */    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */
639      OP_TYPEMINQUERY,   /* 56 */
640    
641      OP_TYPEUPTO,       /* 57 From 0 to n matches */
642      OP_TYPEMINUPTO,    /* 58 */
643      OP_TYPEEXACT,      /* 59 Exactly n matches */
644    
645      OP_TYPEPOSSTAR,    /* 60 Possessified versions */
646      OP_TYPEPOSPLUS,    /* 61 */
647      OP_TYPEPOSQUERY,   /* 62 */
648      OP_TYPEPOSUPTO,    /* 63 */
649    
650      OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */
651      OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */
652      OP_CRPLUS,         /* 66 the minimizing one second. These codes must */
653      OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */
654      OP_CRQUERY,        /* 68 These are for character classes and back refs */
655      OP_CRMINQUERY,     /* 69 */
656      OP_CRRANGE,        /* 70 These are different to the three sets above. */
657      OP_CRMINRANGE,     /* 71 */
658    
659    OP_CLASS,          /* 59 Match a character class, chars < 256 only */    OP_CLASS,          /* 72 Match a character class, chars < 256 only */
660    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative
661                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
662                             character > 255 is encountered. */                             character > 255 is encountered. */
663    
664    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the
665                             class. This does both positive and negative. */                             class. This does both positive and negative. */
666    
667    OP_REF,            /* 62 Match a back reference */    OP_REF,            /* 75 Match a back reference */
668    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */
669    OP_CALLOUT,        /* 64 Call out to external function if provided */    OP_CALLOUT,        /* 77 Call out to external function if provided */
670    
671    OP_ALT,            /* 65 Start of alternation */    OP_ALT,            /* 78 Start of alternation */
672    OP_KET,            /* 66 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 79 End of group that doesn't have an unbounded repeat */
673    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_KETRMAX,        /* 80 These two must remain together and in this */
674    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */
675    
676    /* The assertions must come before ONCE and COND */    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
677    
678    OP_ASSERT,         /* 69 Positive lookahead */    OP_ASSERT,         /* 82 Positive lookahead */
679    OP_ASSERT_NOT,     /* 70 Negative lookahead */    OP_ASSERT_NOT,     /* 83 Negative lookahead */
680    OP_ASSERTBACK,     /* 71 Positive lookbehind */    OP_ASSERTBACK,     /* 84 Positive lookbehind */
681    OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */
682    OP_REVERSE,        /* 73 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */
683    
684    /* 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,
685    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. */
686    
687    OP_ONCE,           /* 74 Once matched, don't back up into the subpattern */    OP_ONCE,           /* 87 Atomic group */
688    OP_COND,           /* 75 Conditional group */    OP_BRA,            /* 88 Start of non-capturing bracket */
689    OP_CREF,           /* 76 Used to hold an extraction string number (cond ref) */    OP_CBRA,           /* 89 Start of capturing bracket */
690      OP_COND,           /* 90 Conditional group */
691    OP_BRAZERO,        /* 77 These two must remain together and in this */  
692    OP_BRAMINZERO,     /* 78 order. */    /* These three must follow the previous three, in the same order. There's a
693      check for >= SBRA to distinguish the two sets. */
694    OP_BRANUMBER,      /* 79 Used for extracting brackets whose number is greater  
695                             than can fit into an opcode. */    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */
696      OP_SCBRA,          /* 92 Start of capturing bracket, check empty */
697    OP_BRA             /* 80 This and greater values are used for brackets that    OP_SCOND,          /* 93 Conditional group, check empty */
698                             extract substrings up to EXTRACT_BASIC_MAX. After  
699                             that, use is made of OP_BRANUMBER. */    OP_CREF,           /* 94 Used to hold a capture number as condition */
700  };    OP_RREF,           /* 95 Used to hold a recursion number as condition */
701      OP_DEF,            /* 96 The DEFINE condition */
 /* 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. */  
702    
703  #define EXTRACT_BASIC_MAX  100    OP_BRAZERO,        /* 97 These two must remain together and in this */
704      OP_BRAMINZERO      /* 98 order. */
705    };
706    
707    
708  /* 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
# Line 608  for debugging. The macro is referenced o Line 711  for debugging. The macro is referenced o
711  #define OP_NAME_LIST \  #define OP_NAME_LIST \
712    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \
713    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \
714    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "anynl", "extuni",                           \
715    "\\Z", "\\z",                                                   \    "\\Z", "\\z",                                                   \
716    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
717    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
718      "*+","++", "?+", "{",                                           \
719    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
720      "*+","++", "?+", "{",                                           \
721    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
722      "*+","++", "?+", "{",                                           \
723    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
724    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
725    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
726    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "AssertB", "AssertB not", "Reverse",                            \
727    "Brazero", "Braminzero", "Branumber", "Bra"    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \
728      "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"
729    
730    
731  /* 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 634  in UTF-8 mode. The code that uses this t Line 741  in UTF-8 mode. The code that uses this t
741    1,                             /* End                                    */ \    1,                             /* End                                    */ \
742    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, 1, 1, 1, 1, 1,  /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \
743    1, 1,                          /* Any, Anybyte                           */ \    1, 1,                          /* Any, Anybyte                           */ \
744    2, 2, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \
745    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
746    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
747    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 642  in UTF-8 mode. The code that uses this t Line 749  in UTF-8 mode. The code that uses this t
749    /* Positive single-char repeats                            ** These are  */ \    /* Positive single-char repeats                            ** These are  */ \
750    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
751    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
752      2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
753    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
754    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
755    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
756      2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
757    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
758    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
759    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
760      2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
761    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
762    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
763    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
# Line 666  in UTF-8 mode. The code that uses this t Line 776  in UTF-8 mode. The code that uses this t
776    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
777    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
778    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* Reverse                                */ \
779    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
780      1+LINK_SIZE,                   /* BRA                                    */ \
781      3+LINK_SIZE,                   /* CBRA                                   */ \
782    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
783      1+LINK_SIZE,                   /* SBRA                                   */ \
784      3+LINK_SIZE,                   /* SCBRA                                  */ \
785      1+LINK_SIZE,                   /* SCOND                                  */ \
786    3,                             /* CREF                                   */ \    3,                             /* CREF                                   */ \
787      3,                             /* RREF                                   */ \
788      1,                             /* DEF                                    */ \
789    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
   3,                             /* BRANUMBER                              */ \  
   1+LINK_SIZE                    /* BRA                                    */ \  
790    
791    
792  /* A magic value for OP_CREF to indicate the "in recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
793    
794  #define CREF_RECURSE  0xffff  #define RREF_ANY  0xffff
795    
796  /* 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
797  tracked. */  tracked. */
# Line 685  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 800  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
800         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
801         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
802         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
803         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
804           ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };
805    
806  /* 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
807  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 740  typedef struct compile_data { Line 856  typedef struct compile_data {
856    const uschar *fcc;            /* Points to case-flipping table */    const uschar *fcc;            /* Points to case-flipping table */
857    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
858    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
859      const uschar *start_workspace;/* The start of working space */
860    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
861    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
862      const uschar *end_pattern;    /* The end of the pattern */
863      uschar *hwm;                  /* High watermark of workspace */
864    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
865    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
866    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
867      int  bracount;                /* Count of capturing parens */
868    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
869    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
870      int  external_options;        /* External (initial) options */
871    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
872    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL nopartial;               /* Set TRUE if partial won't work */
873      int  nltype;                  /* Newline type */
874      int  nllen;                   /* Newline string length */
875      uschar nl[4];                 /* Newline string when fixed length */
876  } compile_data;  } compile_data;
877    
878  /* 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 890  typedef struct recursion_info {
890    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
891    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
892    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 */
893    const uschar *save_start;     /* Old value of md->start_match */    USPTR save_start;             /* Old value of md->start_match */
894    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
895    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
896  } recursion_info;  } recursion_info;
897    
898  /* When compiling in a mode that doesn't use recursive calls to match(),  /* When compiling in a mode that doesn't use recursive calls to match(),
899  a structure is used to remember local variables on the heap. It is defined in  a structure is used to remember local variables on the heap. It is defined in
900  pcre.c, close to the match() function, so that it is easy to keep it in step  pcre_exec.c, close to the match() function, so that it is easy to keep it in
901  with any changes of local variable. However, the pointer to the current frame  step with any changes of local variable. However, the pointer to the current
902  must be saved in some "static" place over a longjmp(). We declare the  frame must be saved in some "static" place over a longjmp(). We declare the
903  structure here so that we can put a pointer in the match_data structure.  structure here so that we can put a pointer in the match_data structure. NOTE:
904  NOTE: This isn't used for a "normal" compilation of pcre. */  This isn't used for a "normal" compilation of pcre. */
905    
906  struct heapframe;  struct heapframe;
907    
908    /* Structure for building a chain of data for holding the values of the subject
909    pointer at the start of each subpattern, so as to detect when an empty string
910    has been matched by a subpattern - to break infinite loops. */
911    
912    typedef struct eptrblock {
913      struct eptrblock *epb_prev;
914      USPTR epb_saved_eptr;
915    } eptrblock;
916    
917    
918  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
919  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
920    
921  typedef struct match_data {  typedef struct match_data {
922    unsigned long int match_call_count; /* As it says */    unsigned long int match_call_count;      /* As it says */
923    unsigned long int match_limit;/* As it says */    unsigned long int match_limit;           /* As it says */
924      unsigned long int match_limit_recursion; /* As it says */
925    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
926    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
927    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
928      int    nltype;                /* Newline type */
929      int    nllen;                 /* Newline string length */
930      uschar nl[4];                 /* Newline string when fixed */
931    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
932    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
933    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
# Line 801  typedef struct match_data { Line 939  typedef struct match_data {
939    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
940    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
941    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
942    const uschar *start_subject;  /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
943    const uschar *end_subject;    /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
944    const uschar *start_match;    /* Start of this match attempt */    USPTR  start_match;           /* Start of this match attempt */
945    const uschar *end_match_ptr;  /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
946    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
947    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
948    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
949      eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
950      int    eptrn;                 /* Next free eptrblock */
951    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
952    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
953    struct heapframe *thisframe;  /* Used only when compiling for no recursion */    struct heapframe *thisframe;  /* Used only when compiling for no recursion */
# Line 823  typedef struct dfa_match_data { Line 963  typedef struct dfa_match_data {
963    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
964    int   moptions;               /* Match options */    int   moptions;               /* Match options */
965    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
966      int    nltype;                /* Newline type */
967      int    nllen;                 /* Newline string length */
968      uschar nl[4];                 /* Newline string when fixed */
969    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
970  } dfa_match_data;  } dfa_match_data;
971    
# Line 859  total length. */ Line 1002  total length. */
1002  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
1003  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1004    
1005  /* 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
1006  pcre_ucp_findchar(). */  codes. */
1007    
1008  typedef struct {  typedef struct {
1009    const char *name;    const char *name;
1010    int value;    pcre_uint16 type;
1011      pcre_uint16 value;
1012  } ucp_type_table;  } ucp_type_table;
1013    
1014    
# Line 892  extern const uschar _pcre_OP_lengths[]; Line 1036  extern const uschar _pcre_OP_lengths[];
1036  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
1037  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1038    
1039  extern int         _pcre_ord2utf8(int, uschar *);  extern BOOL         _pcre_is_newline(const uschar *, const uschar *, int *,
1040  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,                        BOOL);
1041                       const pcre_study_data *, pcre_study_data *);  extern int          _pcre_ord2utf8(int, uschar *);
1042  extern int         _pcre_ucp_findchar(const int, int *, int *);  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1043  extern int         _pcre_valid_utf8(const uschar *, int);                        const pcre_study_data *, pcre_study_data *);
1044  extern BOOL        _pcre_xclass(int, const uschar *);  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);
1045    extern unsigned int _pcre_ucp_othercase(const unsigned int);
1046    extern int          _pcre_valid_utf8(const uschar *, int);
1047    extern BOOL         _pcre_was_newline(const uschar *, const uschar *, int *,
1048                          BOOL);
1049    extern BOOL         _pcre_xclass(int, const uschar *);
1050    
1051  #endif  #endif
1052    

Legend:
Removed from v.85  
changed lines
  Added in v.118

  ViewVC Help
Powered by ViewVC 1.1.5