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

Diff of /code/trunk/pcre_internal.h

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

revision 87 by nigel, Sat Feb 24 21:41:21 2007 UTC revision 117 by ph10, Fri Mar 9 15:59:06 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) 1997-2006 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  /* 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()  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  to process subject strings that are discontinuous by using a smart pointer
# Line 142  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 160  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(void *d, const void *s, size_t n)
214  {  {
215  size_t i;  size_t i;
216    unsigned char *dest = (unsigned char *)d;
217    const unsigned char *src = (const unsigned char *)s;
218  dest += n;  dest += n;
219  src += n;  src += n;
220  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  for (i = 0; i < n; ++i) *(--dest) = *(--src);
221  return dest;  return (void *)dest;
222  }  }
223  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
224  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 270  we know we are in UTF-8 mode. */ Line 321  we know we are in UTF-8 mode. */
321    
322  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
323    c = *eptr; \    c = *eptr; \
324    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
325      { \      { \
326      int gcii; \      int gcii; \
327      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 288  pointer. */ Line 339  pointer. */
339    
340  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
341    c = *eptr; \    c = *eptr; \
342    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
343      { \      { \
344      int gcii; \      int gcii; \
345      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 306  know we are in UTF-8 mode. */ Line 357  know we are in UTF-8 mode. */
357    
358  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
359    c = *eptr++; \    c = *eptr++; \
360    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
361      { \      { \
362      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
363      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 322  know we are in UTF-8 mode. */ Line 373  know we are in UTF-8 mode. */
373    
374  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
375    c = *eptr++; \    c = *eptr++; \
376    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
377      { \      { \
378      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
379      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 339  if there are extra bytes. This is called Line 390  if there are extra bytes. This is called
390    
391  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
392    c = *eptr; \    c = *eptr; \
393    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
394      { \      { \
395      int gcii; \      int gcii; \
396      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 373  Standard C system should have one. */ Line 424  Standard C system should have one. */
424    
425  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
426    
427  /* 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.
428  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
429  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
430  significant end. Make sure they don't overlap! */  out, there is a dummy word in the structure that could be used for the private
431    bits. */
432    
433    #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */
434  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */
435  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */
436  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */
437  #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 */  
438    
439  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
440    
# Line 391  significant end. Make sure they don't ov Line 443  significant end. Make sure they don't ov
443  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
444  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
445    
446    #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY)
447    
448  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
449    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
450     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
451     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| \
452       PCRE_DUPNAMES|PCRE_NEWLINE_BITS)
453    
454  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
455    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
456     PCRE_PARTIAL)     PCRE_PARTIAL|PCRE_NEWLINE_BITS)
457    
458  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
459    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
460     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)
461    
462  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
463    
# Line 434  typedef int BOOL; Line 489  typedef int BOOL;
489  #define FALSE   0  #define FALSE   0
490  #define TRUE    1  #define TRUE    1
491    
492  /* 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). */  
493    
494  #ifndef ESC_e  #ifndef ESC_e
495  #define ESC_e 27  #define ESC_e 27
# Line 447  ESC_n is defined as yet another macro, w Line 500  ESC_n is defined as yet another macro, w
500  #endif  #endif
501    
502  #ifndef ESC_n  #ifndef ESC_n
503  #define ESC_n NEWLINE  #define ESC_n '\n'
504  #endif  #endif
505    
506  #ifndef ESC_r  #ifndef ESC_r
# Line 486  value such as \n. They must have non-zer Line 539  value such as \n. They must have non-zer
539  their negation. Also, they must appear in the same order as in the opcode  their negation. Also, they must appear in the same order as in the opcode
540  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  definitions below, up to ESC_z. There's a dummy for OP_ANY because it
541  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence. The final one must be
542  ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).
543  tests in the code for an escape greater than ESC_b and less than ESC_Z to  There are two tests in the code for an escape greater than ESC_b and less than
544  detect the types that may be repeated. These are the types that consume  ESC_Z to detect the types that may be repeated. These are the types that
545  characters. If any new escapes are put in between that don't consume a  consume characters. If any new escapes are put in between that don't consume a
546  character, that code will have to change. */  character, that code will have to change. */
547    
548  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,
549         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E,         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,
550         ESC_Q, ESC_REF };         ESC_E, ESC_Q, ESC_k, ESC_REF };
551    
552    
553  /* 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
554  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
555  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.
556  Note that whenever this list is updated, the two macro definitions that follow  
557  must also be updated to match. */  To keep stored, compiled patterns compatible, new opcodes should be added
558    immediately before OP_BRA, where (since release 7.0) a gap is left for this
559    purpose.
560    
561    *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
562    that follow must also be updated to match. There is also a table called
563    "coptable" in pcre_dfa_exec.c that must be updated. */
564    
565  enum {  enum {
566    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 521  enum { Line 581  enum {
581    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 */
582    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_NOTPROP,        /* 13 \P (not Unicode property) */
583    OP_PROP,           /* 14 \p (Unicode property) */    OP_PROP,           /* 14 \p (Unicode property) */
584    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_ANYNL,          /* 15 \R (any newline sequence) */
585    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */
586    OP_EOD,            /* 17 End of data: \z */    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */
587      OP_EOD,            /* 18 End of data: \z */
588    OP_OPT,            /* 18 Set runtime options */  
589    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_OPT,            /* 19 Set runtime options */
590    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_CIRC,           /* 20 Start of line - varies with multiline switch */
591    OP_CHAR,           /* 21 Match one character, casefully */    OP_DOLL,           /* 21 End of line - varies with multiline switch */
592    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_CHAR,           /* 22 Match one character, casefully */
593    OP_NOT,            /* 23 Match anything but the following char */    OP_CHARNC,         /* 23 Match one character, caselessly */
594      OP_NOT,            /* 24 Match one character, not the following one */
595    OP_STAR,           /* 24 The maximizing and minimizing versions of */  
596    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_STAR,           /* 25 The maximizing and minimizing versions of */
597    OP_PLUS,           /* 26 the minimizing one second. */    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */
598    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_PLUS,           /* 27 the minimizing one second. */
599    OP_QUERY,          /* 28 */    OP_MINPLUS,        /* 28 This first set applies to single characters.*/
600    OP_MINQUERY,       /* 29 */    OP_QUERY,          /* 29 */
601    OP_UPTO,           /* 30 From 0 to n matches */    OP_MINQUERY,       /* 30 */
602    OP_MINUPTO,        /* 31 */  
603    OP_EXACT,          /* 32 Exactly n matches */    OP_UPTO,           /* 31 From 0 to n matches */
604      OP_MINUPTO,        /* 32 */
605    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    OP_EXACT,          /* 33 Exactly n matches */
606    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */  
607    OP_NOTPLUS,        /* 35 the minimizing one second. */    OP_POSSTAR,        /* 34 Possessified star */
608    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */    OP_POSPLUS,        /* 35 Possessified plus */
609    OP_NOTQUERY,       /* 37 */    OP_POSQUERY,       /* 36 Posesssified query */
610    OP_NOTMINQUERY,    /* 38 */    OP_POSUPTO,        /* 37 Possessified upto */
611    OP_NOTUPTO,        /* 39 From 0 to n matches */  
612    OP_NOTMINUPTO,     /* 40 */    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */
613    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */
614      OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */
615    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */
616    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */
617    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */    OP_NOTMINQUERY,    /* 43 */
618    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */  
619    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_NOTUPTO,        /* 44 From 0 to n matches */
620    OP_TYPEMINQUERY,   /* 47 */    OP_NOTMINUPTO,     /* 45 */
621    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_NOTEXACT,       /* 46 Exactly n matches */
622    OP_TYPEMINUPTO,    /* 49 */  
623    OP_TYPEEXACT,      /* 50 Exactly n matches */    OP_NOTPOSSTAR,     /* 47 Possessified versions */
624      OP_NOTPOSPLUS,     /* 48 */
625    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */    OP_NOTPOSQUERY,    /* 49 */
626    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    OP_NOTPOSUPTO,     /* 50 */
627    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */  
628    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */
629    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */
630    OP_CRMINQUERY,     /* 56 */    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */
631    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */
632    OP_CRMINRANGE,     /* 58 */    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */
633      OP_TYPEMINQUERY,   /* 56 */
634    
635      OP_TYPEUPTO,       /* 57 From 0 to n matches */
636      OP_TYPEMINUPTO,    /* 58 */
637      OP_TYPEEXACT,      /* 59 Exactly n matches */
638    
639      OP_TYPEPOSSTAR,    /* 60 Possessified versions */
640      OP_TYPEPOSPLUS,    /* 61 */
641      OP_TYPEPOSQUERY,   /* 62 */
642      OP_TYPEPOSUPTO,    /* 63 */
643    
644      OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */
645      OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */
646      OP_CRPLUS,         /* 66 the minimizing one second. These codes must */
647      OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */
648      OP_CRQUERY,        /* 68 These are for character classes and back refs */
649      OP_CRMINQUERY,     /* 69 */
650      OP_CRRANGE,        /* 70 These are different to the three sets above. */
651      OP_CRMINRANGE,     /* 71 */
652    
653    OP_CLASS,          /* 59 Match a character class, chars < 256 only */    OP_CLASS,          /* 72 Match a character class, chars < 256 only */
654    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative
655                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
656                             character > 255 is encountered. */                             character > 255 is encountered. */
657    
658    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the
659                             class. This does both positive and negative. */                             class. This does both positive and negative. */
660    
661    OP_REF,            /* 62 Match a back reference */    OP_REF,            /* 75 Match a back reference */
662    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */
663    OP_CALLOUT,        /* 64 Call out to external function if provided */    OP_CALLOUT,        /* 77 Call out to external function if provided */
664    
665    OP_ALT,            /* 65 Start of alternation */    OP_ALT,            /* 78 Start of alternation */
666    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 */
667    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_KETRMAX,        /* 80 These two must remain together and in this */
668    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */
669    
670    /* The assertions must come before ONCE and COND */    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
671    
672    OP_ASSERT,         /* 69 Positive lookahead */    OP_ASSERT,         /* 82 Positive lookahead */
673    OP_ASSERT_NOT,     /* 70 Negative lookahead */    OP_ASSERT_NOT,     /* 83 Negative lookahead */
674    OP_ASSERTBACK,     /* 71 Positive lookbehind */    OP_ASSERTBACK,     /* 84 Positive lookbehind */
675    OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */
676    OP_REVERSE,        /* 73 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */
677    
678    /* 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,
679    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. */
680    
681    OP_ONCE,           /* 74 Once matched, don't back up into the subpattern */    OP_ONCE,           /* 87 Atomic group */
682    OP_COND,           /* 75 Conditional group */    OP_BRA,            /* 88 Start of non-capturing bracket */
683    OP_CREF,           /* 76 Used to hold an extraction string number (cond ref) */    OP_CBRA,           /* 89 Start of capturing bracket */
684      OP_COND,           /* 90 Conditional group */
685    OP_BRAZERO,        /* 77 These two must remain together and in this */  
686    OP_BRAMINZERO,     /* 78 order. */    /* These three must follow the previous three, in the same order. There's a
687      check for >= SBRA to distinguish the two sets. */
688    OP_BRANUMBER,      /* 79 Used for extracting brackets whose number is greater  
689                             than can fit into an opcode. */    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */
690      OP_SCBRA,          /* 92 Start of capturing bracket, check empty */
691    OP_BRA             /* 80 This and greater values are used for brackets that    OP_SCOND,          /* 93 Conditional group, check empty */
692                             extract substrings up to EXTRACT_BASIC_MAX. After  
693                             that, use is made of OP_BRANUMBER. */    OP_CREF,           /* 94 Used to hold a capture number as condition */
694  };    OP_RREF,           /* 95 Used to hold a recursion number as condition */
695      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. */  
696    
697  #define EXTRACT_BASIC_MAX  100    OP_BRAZERO,        /* 97 These two must remain together and in this */
698      OP_BRAMINZERO      /* 98 order. */
699    };
700    
701    
702  /* 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 633  for debugging. The macro is referenced o Line 705  for debugging. The macro is referenced o
705  #define OP_NAME_LIST \  #define OP_NAME_LIST \
706    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \
707    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \
708    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "anynl", "extuni",                           \
709    "\\Z", "\\z",                                                   \    "\\Z", "\\z",                                                   \
710    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
711    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
712      "*+","++", "?+", "{",                                           \
713    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
714      "*+","++", "?+", "{",                                           \
715    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
716      "*+","++", "?+", "{",                                           \
717    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
718    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
719    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
720    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "AssertB", "AssertB not", "Reverse",                            \
721    "Brazero", "Braminzero", "Branumber", "Bra"    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \
722      "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"
723    
724    
725  /* 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 659  in UTF-8 mode. The code that uses this t Line 735  in UTF-8 mode. The code that uses this t
735    1,                             /* End                                    */ \    1,                             /* End                                    */ \
736    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 */ \
737    1, 1,                          /* Any, Anybyte                           */ \    1, 1,                          /* Any, Anybyte                           */ \
738    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \
739    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
740    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
741    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 667  in UTF-8 mode. The code that uses this t Line 743  in UTF-8 mode. The code that uses this t
743    /* Positive single-char repeats                            ** These are  */ \    /* Positive single-char repeats                            ** These are  */ \
744    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
745    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
746      2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
747    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
748    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
749    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
750      2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
751    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
752    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
753    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
754      2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
755    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
756    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
757    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
# Line 691  in UTF-8 mode. The code that uses this t Line 770  in UTF-8 mode. The code that uses this t
770    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
771    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
772    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* Reverse                                */ \
773    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
774      1+LINK_SIZE,                   /* BRA                                    */ \
775      3+LINK_SIZE,                   /* CBRA                                   */ \
776    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
777      1+LINK_SIZE,                   /* SBRA                                   */ \
778      3+LINK_SIZE,                   /* SCBRA                                  */ \
779      1+LINK_SIZE,                   /* SCOND                                  */ \
780    3,                             /* CREF                                   */ \    3,                             /* CREF                                   */ \
781      3,                             /* RREF                                   */ \
782      1,                             /* DEF                                    */ \
783    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
   3,                             /* BRANUMBER                              */ \  
   1+LINK_SIZE                    /* BRA                                    */ \  
784    
785    
786  /* A magic value for OP_CREF to indicate the "in recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
787    
788  #define CREF_RECURSE  0xffff  #define RREF_ANY  0xffff
789    
790  /* 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
791  tracked. */  tracked. */
# Line 710  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 794  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
794         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
795         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
796         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
797         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
798           ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };
799    
800  /* 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
801  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 765  typedef struct compile_data { Line 850  typedef struct compile_data {
850    const uschar *fcc;            /* Points to case-flipping table */    const uschar *fcc;            /* Points to case-flipping table */
851    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
852    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
853      const uschar *start_workspace;/* The start of working space */
854    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
855    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
856      const uschar *end_pattern;    /* The end of the pattern */
857      uschar *hwm;                  /* High watermark of workspace */
858    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
859    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
860    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
861      int  bracount;                /* Count of capturing parens */
862    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
863    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
864      int  external_options;        /* External (initial) options */
865    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
866    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL nopartial;               /* Set TRUE if partial won't work */
867      int  nltype;                  /* Newline type */
868      int  nllen;                   /* Newline string length */
869      uschar nl[4];                 /* Newline string when fixed length */
870  } compile_data;  } compile_data;
871    
872  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
# Line 798  typedef struct recursion_info { Line 891  typedef struct recursion_info {
891    
892  /* 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(),
893  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
894  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
895  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
896  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
897  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:
898  NOTE: This isn't used for a "normal" compilation of pcre. */  This isn't used for a "normal" compilation of pcre. */
899    
900  struct heapframe;  struct heapframe;
901    
902    /* Structure for building a chain of data for holding the values of the subject
903    pointer at the start of each subpattern, so as to detect when an empty string
904    has been matched by a subpattern - to break infinite loops. */
905    
906    typedef struct eptrblock {
907      struct eptrblock *epb_prev;
908      USPTR epb_saved_eptr;
909    } eptrblock;
910    
911    
912  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
913  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
914    
# Line 816  typedef struct match_data { Line 919  typedef struct match_data {
919    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
920    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
921    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
922      int    nltype;                /* Newline type */
923      int    nllen;                 /* Newline string length */
924      uschar nl[4];                 /* Newline string when fixed */
925    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
926    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
927    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
# Line 834  typedef struct match_data { Line 940  typedef struct match_data {
940    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
941    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
942    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
943      eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
944      int    eptrn;                 /* Next free eptrblock */
945    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
946    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
947    struct heapframe *thisframe;  /* Used only when compiling for no recursion */    struct heapframe *thisframe;  /* Used only when compiling for no recursion */
# Line 849  typedef struct dfa_match_data { Line 957  typedef struct dfa_match_data {
957    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
958    int   moptions;               /* Match options */    int   moptions;               /* Match options */
959    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
960      int    nltype;                /* Newline type */
961      int    nllen;                 /* Newline string length */
962      uschar nl[4];                 /* Newline string when fixed */
963    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
964  } dfa_match_data;  } dfa_match_data;
965    
# Line 919  extern const uschar _pcre_OP_lengths[]; Line 1030  extern const uschar _pcre_OP_lengths[];
1030  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
1031  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1032    
1033  extern int         _pcre_ord2utf8(int, uschar *);  extern BOOL         _pcre_is_newline(const uschar *, const uschar *, int *,
1034  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,                        BOOL);
1035                       const pcre_study_data *, pcre_study_data *);  extern int          _pcre_ord2utf8(int, uschar *);
1036  extern int         _pcre_ucp_findprop(const int, int *, int *);  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1037  extern int         _pcre_ucp_othercase(const int);                        const pcre_study_data *, pcre_study_data *);
1038  extern int         _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);
1039  extern BOOL        _pcre_xclass(int, const uschar *);  extern unsigned int _pcre_ucp_othercase(const unsigned int);
1040    extern int          _pcre_valid_utf8(const uschar *, int);
1041    extern BOOL         _pcre_was_newline(const uschar *, const uschar *, int *,
1042                          BOOL);
1043    extern BOOL         _pcre_xclass(int, const uschar *);
1044    
1045  #endif  #endif
1046    

Legend:
Removed from v.87  
changed lines
  Added in v.117

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12