/[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 77 by nigel, Sat Feb 24 21:40:45 2007 UTC revision 101 by ph10, Tue Mar 6 15:19:44 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) 1997-2006 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 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  modules, but which are not relevant to the exported API. This includes some  modules, but which are not relevant to the exported API. This includes some
43  functions whose names all begin with "_pcre_". */  functions whose names all begin with "_pcre_". */
44    
45    #ifndef PCRE_INTERNAL_H
46    #define PCRE_INTERNAL_H
47    
48  /* Define DEBUG to get debugging output on stdout. */  /* Define DEBUG to get debugging output on stdout. */
49    
50  /****  #if 0
51  #define DEBUG  #define DEBUG
52  ****/  #endif
53    
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 112  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  /* Include the public PCRE header */  /* 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  #include "pcre.h"  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 (copy of) the public ucp header, changing the external name into  /* Include the public PCRE header and the definitions of UCP character property
182  a private one. This does no harm, even if we aren't compiling UCP support. */  values. */
183    
184  #define ucp_findchar _pcre_ucp_findchar  #include "pcre.h"
185  #include "ucp.h"  #include "ucp.h"
186    
187  /* When compiling for use with the Virtual Pascal compiler, these functions  /* When compiling for use with the Virtual Pascal compiler, these functions
# Line 127  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 145  case in PCRE. */ Line 209  case in PCRE. */
209  #if HAVE_BCOPY  #if HAVE_BCOPY
210  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
211  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
212  void *  static void *
213  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)
214  {  {
215  int i;  size_t i;
216  dest += n;  dest += n;
217  src += n;  src += n;
218  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  for (i = 0; i < n; ++i) *(--dest) =  *(--src);
219    return dest;
220  }  }
221  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
222  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 254  we know we are in UTF-8 mode. */ Line 319  we know we are in UTF-8 mode. */
319    
320  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
321    c = *eptr; \    c = *eptr; \
322    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
323      { \      { \
324      int gcii; \      int gcii; \
325      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 272  pointer. */ Line 337  pointer. */
337    
338  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
339    c = *eptr; \    c = *eptr; \
340    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
341      { \      { \
342      int gcii; \      int gcii; \
343      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 290  know we are in UTF-8 mode. */ Line 355  know we are in UTF-8 mode. */
355    
356  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
357    c = *eptr++; \    c = *eptr++; \
358    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
359      { \      { \
360      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
361      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 306  know we are in UTF-8 mode. */ Line 371  know we are in UTF-8 mode. */
371    
372  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
373    c = *eptr++; \    c = *eptr++; \
374    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
375      { \      { \
376      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
377      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 323  if there are extra bytes. This is called Line 388  if there are extra bytes. This is called
388    
389  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
390    c = *eptr; \    c = *eptr; \
391    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
392      { \      { \
393      int gcii; \      int gcii; \
394      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 357  Standard C system should have one. */ Line 422  Standard C system should have one. */
422    
423  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
424    
425  /* 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.
426  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
427  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
428  significant end. Make sure they don't overlap! */  out, there is a dummy word in the structure that could be used for the private
429    bits. */
430    
431    #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */
432  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */
433  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */
434  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */
435  #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 */  
436    
437  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
438    
# Line 375  significant end. Make sure they don't ov Line 441  significant end. Make sure they don't ov
441  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
442  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
443    
444    #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY)
445    
446  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
447    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
448     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
449     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| \
450       PCRE_DUPNAMES|PCRE_NEWLINE_BITS)
451    
452  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
453    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
454     PCRE_PARTIAL)     PCRE_PARTIAL|PCRE_NEWLINE_BITS)
455    
456  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
457    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
458     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)
459    
460  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
461    
# Line 418  typedef int BOOL; Line 487  typedef int BOOL;
487  #define FALSE   0  #define FALSE   0
488  #define TRUE    1  #define TRUE    1
489    
490  /* 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). */  
491    
492  #ifndef ESC_e  #ifndef ESC_e
493  #define ESC_e 27  #define ESC_e 27
# Line 431  ESC_n is defined as yet another macro, w Line 498  ESC_n is defined as yet another macro, w
498  #endif  #endif
499    
500  #ifndef ESC_n  #ifndef ESC_n
501  #define ESC_n NEWLINE  #define ESC_n '\n'
502  #endif  #endif
503    
504  #ifndef ESC_r  #ifndef ESC_r
# Line 445  ESC_n is defined as yet another macro, w Line 512  ESC_n is defined as yet another macro, w
512  #define ESC_tee '\t'  #define ESC_tee '\t'
513  #endif  #endif
514    
515  /* 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. */  
516    
517  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 */
518         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 */
519         ESC_Q, ESC_REF };  #define PT_GC         2    /* General characteristic (e.g. L) */
520    #define PT_PC         3    /* Particular characteristic (e.g. Lu) */
521    #define PT_SC         4    /* Script (e.g. Han) */
522    
523  /* 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
524  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 469  contain UTF-8 characters with values gre Line 529  contain UTF-8 characters with values gre
529  #define XCL_END       0    /* Marks end of individual items */  #define XCL_END       0    /* Marks end of individual items */
530  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */
531  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */
532  #define XCL_PROP      3    /* Unicode property (one property code) follows */  #define XCL_PROP      3    /* Unicode property (2-byte property code follows) */
533  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
534    
535    /* These are escaped items that aren't just an encoding of a particular data
536    value such as \n. They must have non-zero values, as check_escape() returns
537    their negation. Also, they must appear in the same order as in the opcode
538    definitions below, up to ESC_z. There's a dummy for OP_ANY because it
539    corresponds to "." rather than an escape sequence. The final one must be
540    ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).
541    There are two tests in the code for an escape greater than ESC_b and less than
542    ESC_Z to detect the types that may be repeated. These are the types that
543    consume characters. If any new escapes are put in between that don't consume a
544    character, that code will have to change. */
545    
546    enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,
547           ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,
548           ESC_E, ESC_Q, ESC_k, ESC_REF };
549    
550    
551  /* 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
552  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
553  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.
554  Note that whenever this list is updated, the two macro definitions that follow  
555  must also be updated to match. */  To keep stored, compiled patterns compatible, new opcodes should be added
556    immediately before OP_BRA, where (since release 7.0) a gap is left for this
557    purpose.
558    
559    *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
560    that follow must also be updated to match. There is also a table called
561    "coptable" in pcre_dfa_exec.c that must be updated. */
562    
563  enum {  enum {
564    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 498  enum { Line 579  enum {
579    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 */
580    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_NOTPROP,        /* 13 \P (not Unicode property) */
581    OP_PROP,           /* 14 \p (Unicode property) */    OP_PROP,           /* 14 \p (Unicode property) */
582    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_ANYNL,          /* 15 \R (any newline sequence) */
583    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */
584    OP_EOD,            /* 17 End of data: \z */    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */
585      OP_EOD,            /* 18 End of data: \z */
586    OP_OPT,            /* 18 Set runtime options */  
587    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_OPT,            /* 19 Set runtime options */
588    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_CIRC,           /* 20 Start of line - varies with multiline switch */
589    OP_CHAR,           /* 21 Match one character, casefully */    OP_DOLL,           /* 21 End of line - varies with multiline switch */
590    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_CHAR,           /* 22 Match one character, casefully */
591    OP_NOT,            /* 23 Match anything but the following char */    OP_CHARNC,         /* 23 Match one character, caselessly */
592      OP_NOT,            /* 24 Match one character, not the following one */
593    OP_STAR,           /* 24 The maximizing and minimizing versions of */  
594    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_STAR,           /* 25 The maximizing and minimizing versions of */
595    OP_PLUS,           /* 26 the minimizing one second. */    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */
596    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_PLUS,           /* 27 the minimizing one second. */
597    OP_QUERY,          /* 28 */    OP_MINPLUS,        /* 28 This first set applies to single characters.*/
598    OP_MINQUERY,       /* 29 */    OP_QUERY,          /* 29 */
599    OP_UPTO,           /* 30 From 0 to n matches */    OP_MINQUERY,       /* 30 */
600    OP_MINUPTO,        /* 31 */  
601    OP_EXACT,          /* 32 Exactly n matches */    OP_UPTO,           /* 31 From 0 to n matches */
602      OP_MINUPTO,        /* 32 */
603    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    OP_EXACT,          /* 33 Exactly n matches */
604    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */  
605    OP_NOTPLUS,        /* 35 the minimizing one second. */    OP_POSSTAR,        /* 34 Possessified star */
606    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */    OP_POSPLUS,        /* 35 Possessified plus */
607    OP_NOTQUERY,       /* 37 */    OP_POSQUERY,       /* 36 Posesssified query */
608    OP_NOTMINQUERY,    /* 38 */    OP_POSUPTO,        /* 37 Possessified upto */
609    OP_NOTUPTO,        /* 39 From 0 to n matches */  
610    OP_NOTMINUPTO,     /* 40 */    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */
611    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */
612      OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */
613    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */
614    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */
615    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */    OP_NOTMINQUERY,    /* 43 */
616    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */  
617    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_NOTUPTO,        /* 44 From 0 to n matches */
618    OP_TYPEMINQUERY,   /* 47 */    OP_NOTMINUPTO,     /* 45 */
619    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_NOTEXACT,       /* 46 Exactly n matches */
620    OP_TYPEMINUPTO,    /* 49 */  
621    OP_TYPEEXACT,      /* 50 Exactly n matches */    OP_NOTPOSSTAR,     /* 47 Possessified versions */
622      OP_NOTPOSPLUS,     /* 48 */
623    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */    OP_NOTPOSQUERY,    /* 49 */
624    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    OP_NOTPOSUPTO,     /* 50 */
625    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */  
626    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */
627    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */
628    OP_CRMINQUERY,     /* 56 */    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */
629    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */
630    OP_CRMINRANGE,     /* 58 */    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */
631      OP_TYPEMINQUERY,   /* 56 */
632    
633      OP_TYPEUPTO,       /* 57 From 0 to n matches */
634      OP_TYPEMINUPTO,    /* 58 */
635      OP_TYPEEXACT,      /* 59 Exactly n matches */
636    
637      OP_TYPEPOSSTAR,    /* 60 Possessified versions */
638      OP_TYPEPOSPLUS,    /* 61 */
639      OP_TYPEPOSQUERY,   /* 62 */
640      OP_TYPEPOSUPTO,    /* 63 */
641    
642      OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */
643      OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */
644      OP_CRPLUS,         /* 66 the minimizing one second. These codes must */
645      OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */
646      OP_CRQUERY,        /* 68 These are for character classes and back refs */
647      OP_CRMINQUERY,     /* 69 */
648      OP_CRRANGE,        /* 70 These are different to the three sets above. */
649      OP_CRMINRANGE,     /* 71 */
650    
651    OP_CLASS,          /* 59 Match a character class, chars < 256 only */    OP_CLASS,          /* 72 Match a character class, chars < 256 only */
652    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative
653                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
654                             character > 255 is encountered. */                             character > 255 is encountered. */
655    
656    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the
657                             class. This does both positive and negative. */                             class. This does both positive and negative. */
658    
659    OP_REF,            /* 62 Match a back reference */    OP_REF,            /* 75 Match a back reference */
660    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */
661    OP_CALLOUT,        /* 64 Call out to external function if provided */    OP_CALLOUT,        /* 77 Call out to external function if provided */
662    
663    OP_ALT,            /* 65 Start of alternation */    OP_ALT,            /* 78 Start of alternation */
664    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 */
665    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_KETRMAX,        /* 80 These two must remain together and in this */
666    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */
667    
668    /* The assertions must come before ONCE and COND */    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
669    
670    OP_ASSERT,         /* 69 Positive lookahead */    OP_ASSERT,         /* 82 Positive lookahead */
671    OP_ASSERT_NOT,     /* 70 Negative lookahead */    OP_ASSERT_NOT,     /* 83 Negative lookahead */
672    OP_ASSERTBACK,     /* 71 Positive lookbehind */    OP_ASSERTBACK,     /* 84 Positive lookbehind */
673    OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */
674    OP_REVERSE,        /* 73 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */
675    
676    /* 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,
677    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. */
678    
679    OP_ONCE,           /* 74 Once matched, don't back up into the subpattern */    OP_ONCE,           /* 87 Atomic group */
680    OP_COND,           /* 75 Conditional group */    OP_BRA,            /* 88 Start of non-capturing bracket */
681    OP_CREF,           /* 76 Used to hold an extraction string number (cond ref) */    OP_CBRA,           /* 89 Start of capturing bracket */
682      OP_COND,           /* 90 Conditional group */
683    OP_BRAZERO,        /* 77 These two must remain together and in this */  
684    OP_BRAMINZERO,     /* 78 order. */    /* These three must follow the previous three, in the same order. There's a
685      check for >= SBRA to distinguish the two sets. */
686    OP_BRANUMBER,      /* 79 Used for extracting brackets whose number is greater  
687                             than can fit into an opcode. */    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */
688      OP_SCBRA,          /* 92 Start of capturing bracket, check empty */
689    OP_BRA             /* 80 This and greater values are used for brackets that    OP_SCOND,          /* 93 Conditional group, check empty */
690                             extract substrings up to EXTRACT_BASIC_MAX. After  
691                             that, use is made of OP_BRANUMBER. */    OP_CREF,           /* 94 Used to hold a capture number as condition */
692  };    OP_RREF,           /* 95 Used to hold a recursion number as condition */
693      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. */  
694    
695  #define EXTRACT_BASIC_MAX  100    OP_BRAZERO,        /* 97 These two must remain together and in this */
696      OP_BRAMINZERO      /* 98 order. */
697    };
698    
699    
700  /* 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 610  for debugging. The macro is referenced o Line 703  for debugging. The macro is referenced o
703  #define OP_NAME_LIST \  #define OP_NAME_LIST \
704    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \
705    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \
706    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "anynl", "extuni",                           \
707    "\\Z", "\\z",                                                   \    "\\Z", "\\z",                                                   \
708    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
709    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
710      "*+","++", "?+", "{",                                           \
711    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
712      "*+","++", "?+", "{",                                           \
713    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
714      "*+","++", "?+", "{",                                           \
715    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
716    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
717    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
718    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "AssertB", "AssertB not", "Reverse",                            \
719    "Brazero", "Braminzero", "Branumber", "Bra"    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \
720      "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"
721    
722    
723  /* 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 636  in UTF-8 mode. The code that uses this t Line 733  in UTF-8 mode. The code that uses this t
733    1,                             /* End                                    */ \    1,                             /* End                                    */ \
734    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 */ \
735    1, 1,                          /* Any, Anybyte                           */ \    1, 1,                          /* Any, Anybyte                           */ \
736    2, 2, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \
737    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
738    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
739    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 644  in UTF-8 mode. The code that uses this t Line 741  in UTF-8 mode. The code that uses this t
741    /* Positive single-char repeats                            ** These are  */ \    /* Positive single-char repeats                            ** These are  */ \
742    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
743    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
744      2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
745    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
746    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
747    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
748      2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
749    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
750    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
751    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
752      2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
753    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
754    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
755    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
# Line 668  in UTF-8 mode. The code that uses this t Line 768  in UTF-8 mode. The code that uses this t
768    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
769    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
770    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* Reverse                                */ \
771    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
772      1+LINK_SIZE,                   /* BRA                                    */ \
773      3+LINK_SIZE,                   /* CBRA                                   */ \
774    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
775      1+LINK_SIZE,                   /* SBRA                                   */ \
776      3+LINK_SIZE,                   /* SCBRA                                  */ \
777      1+LINK_SIZE,                   /* SCOND                                  */ \
778    3,                             /* CREF                                   */ \    3,                             /* CREF                                   */ \
779      3,                             /* RREF                                   */ \
780      1,                             /* DEF                                    */ \
781    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
   3,                             /* BRANUMBER                              */ \  
   1+LINK_SIZE                    /* BRA                                    */ \  
782    
783    
784  /* A magic value for OP_CREF to indicate the "in recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
785    
786  #define CREF_RECURSE  0xffff  #define RREF_ANY  0xffff
787    
788  /* 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
789  tracked. */  tracked. */
# Line 687  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 792  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
792         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
793         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
794         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
795         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
796           ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };
797    
798  /* 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
799  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 742  typedef struct compile_data { Line 848  typedef struct compile_data {
848    const uschar *fcc;            /* Points to case-flipping table */    const uschar *fcc;            /* Points to case-flipping table */
849    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
850    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
851      const uschar *start_workspace;/* The start of working space */
852    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
853    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
854      const uschar *end_pattern;    /* The end of the pattern */
855      uschar *hwm;                  /* High watermark of workspace */
856    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
857    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
858    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
859      int  bracount;                /* Count of capturing parens */
860    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
861    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
862      int  external_options;        /* External (initial) options */
863    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
864    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL nopartial;               /* Set TRUE if partial won't work */
865      int  nltype;                  /* Newline type */
866      int  nllen;                   /* Newline string length */
867      uschar nl[4];                 /* Newline string when fixed length */
868  } compile_data;  } compile_data;
869    
870  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
# Line 768  typedef struct recursion_info { Line 882  typedef struct recursion_info {
882    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
883    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
884    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 */
885    const uschar *save_start;     /* Old value of md->start_match */    USPTR save_start;             /* Old value of md->start_match */
886    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
887    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
888  } recursion_info;  } recursion_info;
889    
890  /* 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(),
891  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
892  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
893  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
894  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
895  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:
896  NOTE: This isn't used for a "normal" compilation of pcre. */  This isn't used for a "normal" compilation of pcre. */
897    
898  struct heapframe;  struct heapframe;
899    
900    /* Structure for building a chain of data for holding the values of the subject
901    pointer at the start of each subpattern, so as to detect when an empty string
902    has been matched by a subpattern - to break infinite loops. */
903    
904    typedef struct eptrblock {
905      struct eptrblock *epb_prev;
906      USPTR epb_saved_eptr;
907    } eptrblock;
908    
909    
910  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
911  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
912    
913  typedef struct match_data {  typedef struct match_data {
914    unsigned long int match_call_count; /* As it says */    unsigned long int match_call_count;      /* As it says */
915    unsigned long int match_limit;/* As it says */    unsigned long int match_limit;           /* As it says */
916      unsigned long int match_limit_recursion; /* As it says */
917    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
918    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
919    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
920      int    nltype;                /* Newline type */
921      int    nllen;                 /* Newline string length */
922      uschar nl[4];                 /* Newline string when fixed */
923    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
924    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
925    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
# Line 803  typedef struct match_data { Line 931  typedef struct match_data {
931    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
932    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
933    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
934    const uschar *start_subject;  /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
935    const uschar *end_subject;    /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
936    const uschar *start_match;    /* Start of this match attempt */    USPTR  start_match;           /* Start of this match attempt */
937    const uschar *end_match_ptr;  /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
938    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
939    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
940    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
941      eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
942      int    eptrn;                 /* Next free eptrblock */
943    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
944    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
945    struct heapframe *thisframe;  /* Used only when compiling for no recursion */    struct heapframe *thisframe;  /* Used only when compiling for no recursion */
# Line 825  typedef struct dfa_match_data { Line 955  typedef struct dfa_match_data {
955    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
956    int   moptions;               /* Match options */    int   moptions;               /* Match options */
957    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
958      int    nltype;                /* Newline type */
959      int    nllen;                 /* Newline string length */
960      uschar nl[4];                 /* Newline string when fixed */
961    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
962  } dfa_match_data;  } dfa_match_data;
963    
# Line 861  total length. */ Line 994  total length. */
994  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
995  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
996    
997  /* 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
998  ucp_findchar(). */  codes. */
999    
1000  typedef struct {  typedef struct {
1001    const char *name;    const char *name;
1002    int value;    pcre_uint16 type;
1003      pcre_uint16 value;
1004  } ucp_type_table;  } ucp_type_table;
1005    
1006    
# Line 894  extern const uschar _pcre_OP_lengths[]; Line 1028  extern const uschar _pcre_OP_lengths[];
1028  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
1029  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1030    
1031  extern int         _pcre_ord2utf8(int, uschar *);  extern BOOL         _pcre_is_newline(const uschar *, const uschar *, int *,
1032  extern void        _pcre_printint(pcre *, FILE *);                        BOOL);
1033  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,  extern int          _pcre_ord2utf8(int, uschar *);
1034                       const pcre_study_data *, pcre_study_data *);  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1035  extern int         _pcre_ucp_findchar(const int, int *, int *);                        const pcre_study_data *, pcre_study_data *);
1036  extern int         _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);
1037  extern BOOL        _pcre_xclass(int, const uschar *);  extern unsigned int _pcre_ucp_othercase(const unsigned int);
1038    extern int          _pcre_valid_utf8(const uschar *, int);
1039    extern BOOL         _pcre_was_newline(const uschar *, const uschar *, int *,
1040                          BOOL);
1041    extern BOOL         _pcre_xclass(int, const uschar *);
1042    
1043    #endif
1044    
1045  /* End of pcre_internal.h */  /* End of pcre_internal.h */

Legend:
Removed from v.77  
changed lines
  Added in v.101

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12