/[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 108 by ph10, Wed Mar 7 11:11:23 2007 UTC revision 168 by ph10, Tue May 29 15:18:18 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 83  setjmp and stdarg are used is when NO_RE Line 83  setjmp and stdarg are used is when NO_RE
83  #include <stdlib.h>  #include <stdlib.h>
84  #include <string.h>  #include <string.h>
85    
86  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
87  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
88    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
89    information there, using __declspec(dllexport) without "extern" we have a
90    definition; with "extern" we have a declaration. The settings here override the
91    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
92    which is all that is needed for applications (they just import the symbols). We
93    use:
94    
95      PCRE_EXP_DECL       for declarations
96      PCRE_EXP_DEFN       for definitions of exported functions
97      PCRE_EXP_DATA_DEFN  for definitions of exported variables
98    
99    The reason for the two DEFN macros is that in non-Windows environments, one
100    does not want to have "extern" before variable definitions because it leads to
101    compiler warnings. So we distinguish between functions and variables. In
102    Windows, the two should always be the same.
103    
104    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
105    which is an application, but needs to import this file in order to "peek" at
106    internals, can #include pcre.h first to get an application's-eye view.
107    
108    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
109    special-purpose environments) might want to stick other stuff in front of
110    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
111    PCRE_EXP_DATA_DEFN only if they are not already set. */
112    
113    #ifndef PCRE_EXP_DECL
114    #  ifdef _WIN32
115    #    ifdef DLL_EXPORT
116    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
117    #      define PCRE_EXP_DEFN       __declspec(dllexport)
118    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
119    #    else
120    #      define PCRE_EXP_DECL       extern
121    #      define PCRE_EXP_DEFN
122    #      define PCRE_EXP_DATA_DEFN
123    #    endif
124    #
125    #  else
126    #    ifdef __cplusplus
127    #      define PCRE_EXP_DECL       extern "C"
128    #    else
129    #      define PCRE_EXP_DECL       extern
130    #    endif
131    #    ifndef PCRE_EXP_DEFN
132    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
133    #    endif
134    #    ifndef PCRE_EXP_DATA_DEFN
135    #      define PCRE_EXP_DATA_DEFN
136    #    endif
137    #  endif
138  #endif  #endif
139    
140  /* 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 175  characters only go up to 0x7fffffff (tho
175  #define NOTACHAR 0xffffffff  #define NOTACHAR 0xffffffff
176    
177  /* 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,
178  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
179  newlines. NLBLOCK, PSSTART, and PSEND are defined in the various modules to  testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
180  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
181  string field names are. */  start/end of string field names are. */
182    
183  #define NLTYPE_FIXED   0     /* Newline is a fixed length string */  #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
184  #define NLTYPE_ANY     1     /* Newline is any Unicode line ending */  #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
185    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
186    
187  /* This macro checks for a newline at the given position */  /* This macro checks for a newline at the given position */
188    
189  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
190    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
191      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
192       _pcre_is_newline((p), NLBLOCK->PSEND, &(NLBLOCK->nllen), utf8) \       _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
193      ) \         utf8)) \
194      : \      : \
195      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
196       (p)[0] == NLBLOCK->nl[0] && \       (p)[0] == NLBLOCK->nl[0] && \
# Line 152  string field names are. */ Line 203  string field names are. */
203  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
204    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
205      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
206       _pcre_was_newline((p), NLBLOCK->PSSTART, &(NLBLOCK->nllen), utf8) \       _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
207      ) \         &(NLBLOCK->nllen), utf8)) \
208      : \      : \
209      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
210       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
# Line 178  must begin with PCRE_. */ Line 229  must begin with PCRE_. */
229  #define USPTR const unsigned char *  #define USPTR const unsigned char *
230  #endif  #endif
231    
232    
233    
234  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
235  values. */  values. */
236    
237  #include "pcre.h"  #include <pcre.h>
238  #include "ucp.h"  #include "ucp.h"
239    
240  /* When compiling for use with the Virtual Pascal compiler, these functions  /* When compiling for use with the Virtual Pascal compiler, these functions
# Line 200  option on the command line. */ Line 253  option on the command line. */
253  /* 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(),
254  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
255  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
256  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. */  
257    
258  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
259  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
260  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
261  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
262  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
263  static void *  static void *
# Line 215  pcre_memmove(void *d, const void *s, siz Line 266  pcre_memmove(void *d, const void *s, siz
266  size_t i;  size_t i;
267  unsigned char *dest = (unsigned char *)d;  unsigned char *dest = (unsigned char *)d;
268  const unsigned char *src = (const unsigned char *)s;  const unsigned char *src = (const unsigned char *)s;
269  dest += n;  if (dest > src)
270  src += n;    {
271  for (i = 0; i < n; ++i) *(--dest) = *(--src);    dest += n;
272  return (void *)dest;    src += n;
273      for (i = 0; i < n; ++i) *(--dest) = *(--src);
274      return (void *)dest;
275      }
276    else
277      {
278      for (i = 0; i < n; ++i) *dest++ = *src++;
279      return (void *)(dest - n);
280      }
281  }  }
282  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
283  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 443  bits. */ Line 502  bits. */
502  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
503  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
504    
505  #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| \
506                               PCRE_NEWLINE_ANYCRLF)
507    
508  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
509    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
# Line 545  ESC_Z to detect the types that may be re Line 605  ESC_Z to detect the types that may be re
605  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
606  character, that code will have to change. */  character, that code will have to change. */
607    
608  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,
609         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_X, ESC_Z, ESC_z,
610         ESC_E, ESC_Q, ESC_k, ESC_REF };         ESC_E, ESC_Q, ESC_k, ESC_REF };
611    
612    
# Line 569  enum { Line 629  enum {
629    
630    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
631    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
632    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
633    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
634    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
635    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
636    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
637    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
638    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
639    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
640    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
641    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character */
642    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */
643    OP_PROP,           /* 14 \p (Unicode property) */    OP_NOTPROP,        /* 14 \P (not Unicode property) */
644    OP_ANYNL,          /* 15 \R (any newline sequence) */    OP_PROP,           /* 15 \p (Unicode property) */
645    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */    OP_ANYNL,          /* 16 \R (any newline sequence) */
646    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */    OP_EXTUNI,         /* 17 \X (extended Unicode sequence */
647    OP_EOD,            /* 18 End of data: \z */    OP_EODN,           /* 18 End of data or \n at end of data: \Z. */
648      OP_EOD,            /* 19 End of data: \z */
649    OP_OPT,            /* 19 Set runtime options */  
650    OP_CIRC,           /* 20 Start of line - varies with multiline switch */    OP_OPT,            /* 20 Set runtime options */
651    OP_DOLL,           /* 21 End of line - varies with multiline switch */    OP_CIRC,           /* 21 Start of line - varies with multiline switch */
652    OP_CHAR,           /* 22 Match one character, casefully */    OP_DOLL,           /* 22 End of line - varies with multiline switch */
653    OP_CHARNC,         /* 23 Match one character, caselessly */    OP_CHAR,           /* 23 Match one character, casefully */
654    OP_NOT,            /* 24 Match one character, not the following one */    OP_CHARNC,         /* 24 Match one character, caselessly */
655      OP_NOT,            /* 25 Match one character, not the following one */
656    OP_STAR,           /* 25 The maximizing and minimizing versions of */  
657    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */    OP_STAR,           /* 26 The maximizing and minimizing versions of */
658    OP_PLUS,           /* 27 the minimizing one second. */    OP_MINSTAR,        /* 27 these six opcodes must come in pairs, with */
659    OP_MINPLUS,        /* 28 This first set applies to single characters.*/    OP_PLUS,           /* 28 the minimizing one second. */
660    OP_QUERY,          /* 29 */    OP_MINPLUS,        /* 29 This first set applies to single characters.*/
661    OP_MINQUERY,       /* 30 */    OP_QUERY,          /* 30 */
662      OP_MINQUERY,       /* 31 */
663    OP_UPTO,           /* 31 From 0 to n matches */  
664    OP_MINUPTO,        /* 32 */    OP_UPTO,           /* 32 From 0 to n matches */
665    OP_EXACT,          /* 33 Exactly n matches */    OP_MINUPTO,        /* 33 */
666      OP_EXACT,          /* 34 Exactly n matches */
667    OP_POSSTAR,        /* 34 Possessified star */  
668    OP_POSPLUS,        /* 35 Possessified plus */    OP_POSSTAR,        /* 35 Possessified star */
669    OP_POSQUERY,       /* 36 Posesssified query */    OP_POSPLUS,        /* 36 Possessified plus */
670    OP_POSUPTO,        /* 37 Possessified upto */    OP_POSQUERY,       /* 37 Posesssified query */
671      OP_POSUPTO,        /* 38 Possessified upto */
672    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */  
673    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */    OP_NOTSTAR,        /* 39 The maximizing and minimizing versions of */
674    OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */    OP_NOTMINSTAR,     /* 40 these six opcodes must come in pairs, with */
675    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */    OP_NOTPLUS,        /* 41 the minimizing one second. They must be in */
676    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */    OP_NOTMINPLUS,     /* 42 exactly the same order as those above. */
677    OP_NOTMINQUERY,    /* 43 */    OP_NOTQUERY,       /* 43 This set applies to "not" single characters. */
678      OP_NOTMINQUERY,    /* 44 */
679    OP_NOTUPTO,        /* 44 From 0 to n matches */  
680    OP_NOTMINUPTO,     /* 45 */    OP_NOTUPTO,        /* 45 From 0 to n matches */
681    OP_NOTEXACT,       /* 46 Exactly n matches */    OP_NOTMINUPTO,     /* 46 */
682      OP_NOTEXACT,       /* 47 Exactly n matches */
683    OP_NOTPOSSTAR,     /* 47 Possessified versions */  
684    OP_NOTPOSPLUS,     /* 48 */    OP_NOTPOSSTAR,     /* 48 Possessified versions */
685    OP_NOTPOSQUERY,    /* 49 */    OP_NOTPOSPLUS,     /* 49 */
686    OP_NOTPOSUPTO,     /* 50 */    OP_NOTPOSQUERY,    /* 50 */
687      OP_NOTPOSUPTO,     /* 51 */
688    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */  
689    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */    OP_TYPESTAR,       /* 52 The maximizing and minimizing versions of */
690    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */    OP_TYPEMINSTAR,    /* 53 these six opcodes must come in pairs, with */
691    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */    OP_TYPEPLUS,       /* 54 the minimizing one second. These codes must */
692    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */    OP_TYPEMINPLUS,    /* 55 be in exactly the same order as those above. */
693    OP_TYPEMINQUERY,   /* 56 */    OP_TYPEQUERY,      /* 56 This set applies to character types such as \d */
694      OP_TYPEMINQUERY,   /* 57 */
695    OP_TYPEUPTO,       /* 57 From 0 to n matches */  
696    OP_TYPEMINUPTO,    /* 58 */    OP_TYPEUPTO,       /* 58 From 0 to n matches */
697    OP_TYPEEXACT,      /* 59 Exactly n matches */    OP_TYPEMINUPTO,    /* 59 */
698      OP_TYPEEXACT,      /* 60 Exactly n matches */
699    OP_TYPEPOSSTAR,    /* 60 Possessified versions */  
700    OP_TYPEPOSPLUS,    /* 61 */    OP_TYPEPOSSTAR,    /* 61 Possessified versions */
701    OP_TYPEPOSQUERY,   /* 62 */    OP_TYPEPOSPLUS,    /* 62 */
702    OP_TYPEPOSUPTO,    /* 63 */    OP_TYPEPOSQUERY,   /* 63 */
703      OP_TYPEPOSUPTO,    /* 64 */
704    OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */  
705    OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */    OP_CRSTAR,         /* 65 The maximizing and minimizing versions of */
706    OP_CRPLUS,         /* 66 the minimizing one second. These codes must */    OP_CRMINSTAR,      /* 66 all these opcodes must come in pairs, with */
707    OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */    OP_CRPLUS,         /* 67 the minimizing one second. These codes must */
708    OP_CRQUERY,        /* 68 These are for character classes and back refs */    OP_CRMINPLUS,      /* 68 be in exactly the same order as those above. */
709    OP_CRMINQUERY,     /* 69 */    OP_CRQUERY,        /* 69 These are for character classes and back refs */
710    OP_CRRANGE,        /* 70 These are different to the three sets above. */    OP_CRMINQUERY,     /* 70 */
711    OP_CRMINRANGE,     /* 71 */    OP_CRRANGE,        /* 71 These are different to the three sets above. */
712      OP_CRMINRANGE,     /* 72 */
713    
714    OP_CLASS,          /* 72 Match a character class, chars < 256 only */    OP_CLASS,          /* 73 Match a character class, chars < 256 only */
715    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 74 Same, but the bitmap was created from a negative
716                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
717                             character > 255 is encountered. */                             character > 255 is encountered. */
718    
719    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 75 Extended class for handling UTF-8 chars within the
720                             class. This does both positive and negative. */                             class. This does both positive and negative. */
721    
722    OP_REF,            /* 75 Match a back reference */    OP_REF,            /* 76 Match a back reference */
723    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 77 Match a numbered subpattern (possibly recursive) */
724    OP_CALLOUT,        /* 77 Call out to external function if provided */    OP_CALLOUT,        /* 78 Call out to external function if provided */
725    
726    OP_ALT,            /* 78 Start of alternation */    OP_ALT,            /* 79 Start of alternation */
727    OP_KET,            /* 79 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 80 End of group that doesn't have an unbounded repeat */
728    OP_KETRMAX,        /* 80 These two must remain together and in this */    OP_KETRMAX,        /* 81 These two must remain together and in this */
729    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 82 order. They are for groups the repeat for ever. */
730    
731    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
732    
733    OP_ASSERT,         /* 82 Positive lookahead */    OP_ASSERT,         /* 83 Positive lookahead */
734    OP_ASSERT_NOT,     /* 83 Negative lookahead */    OP_ASSERT_NOT,     /* 84 Negative lookahead */
735    OP_ASSERTBACK,     /* 84 Positive lookbehind */    OP_ASSERTBACK,     /* 85 Positive lookbehind */
736    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 86 Negative lookbehind */
737    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 87 Move pointer back - used in lookbehind assertions */
738    
739    /* 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,
740    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. */
741    
742    OP_ONCE,           /* 87 Atomic group */    OP_ONCE,           /* 88 Atomic group */
743    OP_BRA,            /* 88 Start of non-capturing bracket */    OP_BRA,            /* 89 Start of non-capturing bracket */
744    OP_CBRA,           /* 89 Start of capturing bracket */    OP_CBRA,           /* 90 Start of capturing bracket */
745    OP_COND,           /* 90 Conditional group */    OP_COND,           /* 91 Conditional group */
746    
747    /* 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
748    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
749    
750    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 92 Start of non-capturing bracket, check empty  */
751    OP_SCBRA,          /* 92 Start of capturing bracket, check empty */    OP_SCBRA,          /* 93 Start of capturing bracket, check empty */
752    OP_SCOND,          /* 93 Conditional group, check empty */    OP_SCOND,          /* 94 Conditional group, check empty */
753    
754    OP_CREF,           /* 94 Used to hold a capture number as condition */    OP_CREF,           /* 95 Used to hold a capture number as condition */
755    OP_RREF,           /* 95 Used to hold a recursion number as condition */    OP_RREF,           /* 96 Used to hold a recursion number as condition */
756    OP_DEF,            /* 96 The DEFINE condition */    OP_DEF,            /* 97 The DEFINE condition */
757    
758    OP_BRAZERO,        /* 97 These two must remain together and in this */    OP_BRAZERO,        /* 98 These two must remain together and in this */
759    OP_BRAMINZERO      /* 98 order. */    OP_BRAMINZERO      /* 99 order. */
760  };  };
761    
762    
# Line 703  enum { Line 764  enum {
764  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
765    
766  #define OP_NAME_LIST \  #define OP_NAME_LIST \
767    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
768    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \
769    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "anynl", "extuni",                           \
770    "\\Z", "\\z",                                                   \    "\\Z", "\\z",                                                   \
# Line 733  in UTF-8 mode. The code that uses this t Line 794  in UTF-8 mode. The code that uses this t
794    
795  #define OP_LENGTHS \  #define OP_LENGTHS \
796    1,                             /* End                                    */ \    1,                             /* End                                    */ \
797    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                     */ \
798      1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
799    1, 1,                          /* Any, Anybyte                           */ \    1, 1,                          /* Any, Anybyte                           */ \
800    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \
801    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
# Line 795  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 884  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 935  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 944  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 1030  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.108  
changed lines
  Added in v.168
webmaster@exim.org
 ViewVC Help
Powered by ViewVC 1.1.12