/[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 93 by nigel, Sat Feb 24 21:41:42 2007 UTC revision 199 by ph10, Tue Jul 31 14:39:09 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 67  be absolutely sure we get our version. * Line 67  be absolutely sure we get our version. *
67  #endif  #endif
68    
69    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
70  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
71  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
72    
# Line 83  setjmp and stdarg are used is when NO_RE Line 79  setjmp and stdarg are used is when NO_RE
79  #include <stdlib.h>  #include <stdlib.h>
80  #include <string.h>  #include <string.h>
81    
82  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
83  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
84    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
85    information there, using __declspec(dllexport) without "extern" we have a
86    definition; with "extern" we have a declaration. The settings here override the
87    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
88    which is all that is needed for applications (they just import the symbols). We
89    use:
90    
91      PCRE_EXP_DECL       for declarations
92      PCRE_EXP_DEFN       for definitions of exported functions
93      PCRE_EXP_DATA_DEFN  for definitions of exported variables
94    
95    The reason for the two DEFN macros is that in non-Windows environments, one
96    does not want to have "extern" before variable definitions because it leads to
97    compiler warnings. So we distinguish between functions and variables. In
98    Windows, the two should always be the same.
99    
100    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
101    which is an application, but needs to import this file in order to "peek" at
102    internals, can #include pcre.h first to get an application's-eye view.
103    
104    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
105    special-purpose environments) might want to stick other stuff in front of
106    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
107    PCRE_EXP_DATA_DEFN only if they are not already set. */
108    
109    #ifndef PCRE_EXP_DECL
110    #  ifdef _WIN32
111    #    ifndef PCRE_STATIC
112    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
113    #      define PCRE_EXP_DEFN       __declspec(dllexport)
114    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
115    #    else
116    #      define PCRE_EXP_DECL       extern
117    #      define PCRE_EXP_DEFN
118    #      define PCRE_EXP_DATA_DEFN
119    #    endif
120    #  else
121    #    ifdef __cplusplus
122    #      define PCRE_EXP_DECL       extern "C"
123    #    else
124    #      define PCRE_EXP_DECL       extern
125    #    endif
126    #    ifndef PCRE_EXP_DEFN
127    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
128    #    endif
129    #    ifndef PCRE_EXP_DATA_DEFN
130    #      define PCRE_EXP_DATA_DEFN
131    #    endif
132    #  endif
133  #endif  #endif
134    
135  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
# Line 125  characters only go up to 0x7fffffff (tho Line 170  characters only go up to 0x7fffffff (tho
170  #define NOTACHAR 0xffffffff  #define NOTACHAR 0xffffffff
171    
172  /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,  /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
173  and "all" at present). The following macros are used to package up testing for  "any" and "anycrlf" at present). The following macros are used to package up
174  newlines. NLBLOCK, PSSTART, and PSEND are defined in the various modules to  testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
175  indicate in which datablock the parameters exist, and what the start/end of  modules to indicate in which datablock the parameters exist, and what the
176  string field names are. */  start/end of string field names are. */
177    
178  #define NLTYPE_FIXED   0     /* Newline is a fixed length string */  #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
179  #define NLTYPE_ANY     1     /* Newline is any Unicode line ending */  #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
180    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
181    
182  /* This macro checks for a newline at the given position */  /* This macro checks for a newline at the given position */
183    
184  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
185    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
186      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
187       _pcre_is_newline((p), NLBLOCK->PSEND, &(NLBLOCK->nllen), utf8) \       _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
188      ) \         utf8)) \
189      : \      : \
190      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
191       (p)[0] == NLBLOCK->nl[0] && \       (p)[0] == NLBLOCK->nl[0] && \
# Line 152  string field names are. */ Line 198  string field names are. */
198  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
199    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
200      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
201       _pcre_was_newline((p), NLBLOCK->PSSTART, &(NLBLOCK->nllen), utf8) \       _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
202      ) \         &(NLBLOCK->nllen), utf8)) \
203      : \      : \
204      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
205       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
# Line 178  must begin with PCRE_. */ Line 224  must begin with PCRE_. */
224  #define USPTR const unsigned char *  #define USPTR const unsigned char *
225  #endif  #endif
226    
227    
228    
229  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
230  values. */  values. */
231    
# Line 189  need to have their names changed. PCRE m Line 237  need to have their names changed. PCRE m
237  option on the command line. */  option on the command line. */
238    
239  #ifdef VPCOMPAT  #ifdef VPCOMPAT
240    #define strlen(s)        _strlen(s)
241  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
242    #define memcmp(s,c,n)    _memcmp(s,c,n)
243  #define memcpy(d,s,n)    _memcpy(d,s,n)  #define memcpy(d,s,n)    _memcpy(d,s,n)
244  #define memmove(d,s,n)   _memmove(d,s,n)  #define memmove(d,s,n)   _memmove(d,s,n)
245  #define memset(s,c,n)    _memset(s,c,n)  #define memset(s,c,n)    _memset(s,c,n)
# Line 198  option on the command line. */ Line 248  option on the command line. */
248  /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),  /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
249  define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY  define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
250  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
251  neither (there some non-Unix environments where this is the case). This assumes  neither (there some non-Unix environments where this is the case). */
 that all calls to memmove are moving strings upwards in store, which is the  
 case in PCRE. */  
252    
253  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
254  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
255  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
256  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
257  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
258  static void *  static void *
259  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
260  {  {
261  size_t i;  size_t i;
262  dest += n;  unsigned char *dest = (unsigned char *)d;
263  src += n;  const unsigned char *src = (const unsigned char *)s;
264  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
265  return dest;    {
266      dest += n;
267      src += n;
268      for (i = 0; i < n; ++i) *(--dest) = *(--src);
269      return (void *)dest;
270      }
271    else
272      {
273      for (i = 0; i < n; ++i) *dest++ = *src++;
274      return (void *)(dest - n);
275      }
276  }  }
277  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
278  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 439  bits. */ Line 497  bits. */
497  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
498  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
499    
500  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY)  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
501                               PCRE_NEWLINE_ANYCRLF)
502    
503  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
504    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
# Line 541  ESC_Z to detect the types that may be re Line 600  ESC_Z to detect the types that may be re
600  consume 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
601  character, that code will have to change. */  character, that code will have to change. */
602    
603  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_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
604         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,         ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,
605         ESC_E, ESC_Q, ESC_k, ESC_REF };         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_k, ESC_REF };
606    
607    
608  /* 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
# Line 565  enum { Line 624  enum {
624    
625    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
626    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
627    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
628    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
629    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
630    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
631    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
632    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
633    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
634    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
635    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
636    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character */
637    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */
638    OP_PROP,           /* 14 \p (Unicode property) */    OP_NOTPROP,        /* 14 \P (not Unicode property) */
639    OP_ANYNL,          /* 15 \R (any newline sequence) */    OP_PROP,           /* 15 \p (Unicode property) */
640    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */    OP_ANYNL,          /* 16 \R (any newline sequence) */
641    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */
642    OP_EOD,            /* 18 End of data: \z */    OP_HSPACE,         /* 18 \h (horizontal whitespace) */
643      OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */
644    OP_OPT,            /* 19 Set runtime options */    OP_VSPACE,         /* 20 \v (vertical whitespace) */
645    OP_CIRC,           /* 20 Start of line - varies with multiline switch */    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */
646    OP_DOLL,           /* 21 End of line - varies with multiline switch */    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */
647    OP_CHAR,           /* 22 Match one character, casefully */    OP_EOD,            /* 23 End of data: \z */
648    OP_CHARNC,         /* 23 Match one character, caselessly */  
649    OP_NOT,            /* 24 Match one character, not the following one */    OP_OPT,            /* 24 Set runtime options */
650      OP_CIRC,           /* 25 Start of line - varies with multiline switch */
651    OP_STAR,           /* 25 The maximizing and minimizing versions of */    OP_DOLL,           /* 26 End of line - varies with multiline switch */
652    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */    OP_CHAR,           /* 27 Match one character, casefully */
653    OP_PLUS,           /* 27 the minimizing one second. */    OP_CHARNC,         /* 28 Match one character, caselessly */
654    OP_MINPLUS,        /* 28 This first set applies to single characters.*/    OP_NOT,            /* 29 Match one character, not the following one */
655    OP_QUERY,          /* 29 */  
656    OP_MINQUERY,       /* 30 */    OP_STAR,           /* 30 The maximizing and minimizing versions of */
657      OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */
658    OP_UPTO,           /* 31 From 0 to n matches */    OP_PLUS,           /* 32 the minimizing one second. */
659    OP_MINUPTO,        /* 32 */    OP_MINPLUS,        /* 33 This first set applies to single characters.*/
660    OP_EXACT,          /* 33 Exactly n matches */    OP_QUERY,          /* 34 */
661      OP_MINQUERY,       /* 35 */
662    OP_POSSTAR,        /* 34 Possessified star */  
663    OP_POSPLUS,        /* 35 Possessified plus */    OP_UPTO,           /* 36 From 0 to n matches */
664    OP_POSQUERY,       /* 36 Posesssified query */    OP_MINUPTO,        /* 37 */
665    OP_POSUPTO,        /* 37 Possessified upto */    OP_EXACT,          /* 38 Exactly n matches */
666    
667    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */    OP_POSSTAR,        /* 39 Possessified star */
668    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */    OP_POSPLUS,        /* 40 Possessified plus */
669    OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */    OP_POSQUERY,       /* 41 Posesssified query */
670    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */    OP_POSUPTO,        /* 42 Possessified upto */
671    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */  
672    OP_NOTMINQUERY,    /* 43 */    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */
673      OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */
674    OP_NOTUPTO,        /* 44 From 0 to n matches */    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */
675    OP_NOTMINUPTO,     /* 45 */    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */
676    OP_NOTEXACT,       /* 46 Exactly n matches */    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */
677      OP_NOTMINQUERY,    /* 48 */
678    OP_NOTPOSSTAR,     /* 47 Possessified versions */  
679    OP_NOTPOSPLUS,     /* 48 */    OP_NOTUPTO,        /* 49 From 0 to n matches */
680    OP_NOTPOSQUERY,    /* 49 */    OP_NOTMINUPTO,     /* 50 */
681    OP_NOTPOSUPTO,     /* 50 */    OP_NOTEXACT,       /* 51 Exactly n matches */
682    
683    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */    OP_NOTPOSSTAR,     /* 52 Possessified versions */
684    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */    OP_NOTPOSPLUS,     /* 53 */
685    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */    OP_NOTPOSQUERY,    /* 54 */
686    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */    OP_NOTPOSUPTO,     /* 55 */
687    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */  
688    OP_TYPEMINQUERY,   /* 56 */    OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */
689      OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */
690    OP_TYPEUPTO,       /* 57 From 0 to n matches */    OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */
691    OP_TYPEMINUPTO,    /* 58 */    OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */
692    OP_TYPEEXACT,      /* 59 Exactly n matches */    OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */
693      OP_TYPEMINQUERY,   /* 61 */
694    OP_TYPEPOSSTAR,    /* 60 Possessified versions */  
695    OP_TYPEPOSPLUS,    /* 61 */    OP_TYPEUPTO,       /* 62 From 0 to n matches */
696    OP_TYPEPOSQUERY,   /* 62 */    OP_TYPEMINUPTO,    /* 63 */
697    OP_TYPEPOSUPTO,    /* 63 */    OP_TYPEEXACT,      /* 64 Exactly n matches */
698    
699    OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */    OP_TYPEPOSSTAR,    /* 65 Possessified versions */
700    OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */    OP_TYPEPOSPLUS,    /* 66 */
701    OP_CRPLUS,         /* 66 the minimizing one second. These codes must */    OP_TYPEPOSQUERY,   /* 67 */
702    OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */    OP_TYPEPOSUPTO,    /* 68 */
703    OP_CRQUERY,        /* 68 These are for character classes and back refs */  
704    OP_CRMINQUERY,     /* 69 */    OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */
705    OP_CRRANGE,        /* 70 These are different to the three sets above. */    OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */
706    OP_CRMINRANGE,     /* 71 */    OP_CRPLUS,         /* 71 the minimizing one second. These codes must */
707      OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */
708      OP_CRQUERY,        /* 73 These are for character classes and back refs */
709      OP_CRMINQUERY,     /* 74 */
710      OP_CRRANGE,        /* 75 These are different to the three sets above. */
711      OP_CRMINRANGE,     /* 76 */
712    
713    OP_CLASS,          /* 72 Match a character class, chars < 256 only */    OP_CLASS,          /* 77 Match a character class, chars < 256 only */
714    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative
715                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
716                             character > 255 is encountered. */                             character > 255 is encountered. */
717    
718    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the
719                             class. This does both positive and negative. */                             class. This does both positive and negative. */
720    
721    OP_REF,            /* 75 Match a back reference */    OP_REF,            /* 80 Match a back reference */
722    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */
723    OP_CALLOUT,        /* 77 Call out to external function if provided */    OP_CALLOUT,        /* 82 Call out to external function if provided */
724    
725    OP_ALT,            /* 78 Start of alternation */    OP_ALT,            /* 83 Start of alternation */
726    OP_KET,            /* 79 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */
727    OP_KETRMAX,        /* 80 These two must remain together and in this */    OP_KETRMAX,        /* 85 These two must remain together and in this */
728    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */
729    
730    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
731    
732    OP_ASSERT,         /* 82 Positive lookahead */    OP_ASSERT,         /* 87 Positive lookahead */
733    OP_ASSERT_NOT,     /* 83 Negative lookahead */    OP_ASSERT_NOT,     /* 88 Negative lookahead */
734    OP_ASSERTBACK,     /* 84 Positive lookbehind */    OP_ASSERTBACK,     /* 89 Positive lookbehind */
735    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */
736    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */
737    
738    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
739    as there's 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. */
740    
741    OP_ONCE,           /* 87 Atomic group */    OP_ONCE,           /* 92 Atomic group */
742    OP_BRA,            /* 88 Start of non-capturing bracket */    OP_BRA,            /* 93 Start of non-capturing bracket */
743    OP_CBRA,           /* 89 Start of capturing bracket */    OP_CBRA,           /* 94 Start of capturing bracket */
744    OP_COND,           /* 90 Conditional group */    OP_COND,           /* 95 Conditional group */
745    
746    /* These three must follow the previous three, in the same order. There's a    /* These three must follow the previous three, in the same order. There's a
747    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
748    
749    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */
750    OP_SCBRA,          /* 92 Start of capturing bracket, check empty */    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */
751    OP_SCOND,          /* 93 Conditional group, check empty */    OP_SCOND,          /* 98 Conditional group, check empty */
752    
753    OP_CREF,           /* 94 Used to hold a capture number as condition */    OP_CREF,           /* 99 Used to hold a capture number as condition */
754    OP_RREF,           /* 95 Used to hold a recursion number as condition */    OP_RREF,           /* 100 Used to hold a recursion number as condition */
755    OP_DEF,            /* 96 The DEFINE condition */    OP_DEF,            /* 101 The DEFINE condition */
756    
757    OP_BRAZERO,        /* 97 These two must remain together and in this */    OP_BRAZERO,        /* 102 These two must remain together and in this */
758    OP_BRAMINZERO      /* 98 order. */    OP_BRAMINZERO      /* 103 order. */
759  };  };
760    
761    
# Line 699  enum { Line 763  enum {
763  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
764    
765  #define OP_NAME_LIST \  #define OP_NAME_LIST \
766    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
767    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \
768    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
769    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
770    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
771    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
772    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 714  for debugging. The macro is referenced o Line 778  for debugging. The macro is referenced o
778    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
779    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
780    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
781    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
782    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"
783    
784    
# Line 729  in UTF-8 mode. The code that uses this t Line 793  in UTF-8 mode. The code that uses this t
793    
794  #define OP_LENGTHS \  #define OP_LENGTHS \
795    1,                             /* End                                    */ \    1,                             /* End                                    */ \
796    1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
797      1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
798    1, 1,                          /* Any, Anybyte                           */ \    1, 1,                          /* Any, Anybyte                           */ \
799    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
800      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
801    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
802    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
803    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 791  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 857  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
857         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
858         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
859         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
860         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };
861    
862  /* 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
863  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 880  typedef struct recursion_info { Line 946  typedef struct recursion_info {
946    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
947    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
948    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 */
949    USPTR save_start;             /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
950    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
951    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
952  } recursion_info;  } recursion_info;
953    
 /* When compiling in a mode that doesn't use recursive calls to match(),  
 a structure is used to remember local variables on the heap. It is defined in  
 pcre_exec.c, close to the match() function, so that it is easy to keep it in  
 step with any changes of local variable. However, the pointer to the current  
 frame must be saved in some "static" place over a longjmp(). We declare the  
 structure here so that we can put a pointer in the match_data structure. NOTE:  
 This isn't used for a "normal" compilation of pcre. */  
   
 struct heapframe;  
   
954  /* Structure for building a chain of data for holding the values of the subject  /* Structure for building a chain of data for holding the values of the subject
955  pointer at the start of each subpattern, so as to detect when an empty string  pointer at the start of each subpattern, so as to detect when an empty string
956  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops. */
# Line 931  typedef struct match_data { Line 987  typedef struct match_data {
987    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
988    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
989    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
990    USPTR  start_match;           /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
991    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
992    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
993    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
# Line 940  typedef struct match_data { Line 996  typedef struct match_data {
996    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
997    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
998    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
   struct heapframe *thisframe;  /* Used only when compiling for no recursion */  
999  } match_data;  } match_data;
1000    
1001  /* A similar structure is used for the same purpose by the DFA matching  /* A similar structure is used for the same purpose by the DFA matching
# Line 1026  extern const uschar _pcre_OP_lengths[]; Line 1081  extern const uschar _pcre_OP_lengths[];
1081  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
1082  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1083    
1084  extern BOOL         _pcre_is_newline(const uschar *, const uschar *, int *,  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
1085                        BOOL);                        int *, BOOL);
1086  extern int          _pcre_ord2utf8(int, uschar *);  extern int          _pcre_ord2utf8(int, uschar *);
1087  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1088                        const pcre_study_data *, pcre_study_data *);                        const pcre_study_data *, pcre_study_data *);
1089  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);
1090  extern unsigned int _pcre_ucp_othercase(const unsigned int);  extern unsigned int _pcre_ucp_othercase(const unsigned int);
1091  extern int          _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1092  extern BOOL         _pcre_was_newline(const uschar *, const uschar *, int *,  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1093                        BOOL);                        int *, BOOL);
1094  extern BOOL         _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_xclass(int, const uschar *);
1095    
1096  #endif  #endif

Legend:
Removed from v.93  
changed lines
  Added in v.199

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12