/[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 268 by ph10, Thu Nov 15 10:28:09 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 200  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 *
# Line 215  pcre_memmove(void *d, const void *s, siz Line 261  pcre_memmove(void *d, const void *s, siz
261  size_t i;  size_t i;
262  unsigned char *dest = (unsigned char *)d;  unsigned char *dest = (unsigned char *)d;
263  const unsigned char *src = (const unsigned char *)s;  const unsigned char *src = (const unsigned char *)s;
264  dest += n;  if (dest > src)
265  src += n;    {
266  for (i = 0; i < n; ++i) *(--dest) = *(--src);    dest += n;
267  return (void *)dest;    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 304  capturing parenthesis numbers in back re Line 358  capturing parenthesis numbers in back re
358    
359  /* When UTF-8 encoding is being used, a character is no longer just a single  /* When UTF-8 encoding is being used, a character is no longer just a single
360  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
361  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
362    never be called in byte mode. To make sure it can never even appear when UTF-8
363    support is omitted, we don't even define it. */
364    
365  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
366  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 312  byte-mode, and more complicated ones for Line 368  byte-mode, and more complicated ones for
368  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
369  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
370  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
371  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
372    
373  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
374    
# Line 405  if there are extra bytes. This is called Line 461  if there are extra bytes. This is called
461      }      }
462    
463  /* If the pointer is not at the start of a character, move it back until  /* If the pointer is not at the start of a character, move it back until
464  it is. Called only in UTF-8 mode. */  it is. This is called only in UTF-8 mode - we don't put a test within the macro
465    because almost all calls are already within a block of UTF-8 only code. */
466    
467  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
468    
469  #endif  #endif
470    
# Line 424  Standard C system should have one. */ Line 481  Standard C system should have one. */
481    
482  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
483    
484  /* Private options flags start at the most significant end of the four bytes.  /* Private flags containing information about the compiled regex. They used to
485  The public options defined in pcre.h start at the least significant end. Make  live at the top end of the options word, but that got almost full, so now they
486  sure they don't overlap! The bits are getting a bit scarce now -- when we run  are in a 16-bit flags word. */
487  out, there is a dummy word in the structure that could be used for the private  
488  bits. */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
489    #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
490  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
491  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
492  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
493  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
 #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */  
494    
495  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
496    
# Line 443  bits. */ Line 499  bits. */
499  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
500  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
501    
502  #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| \
503                               PCRE_NEWLINE_ANYCRLF)
504    
505  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
506    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
507     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
508     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| \
509     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
510    
511  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
512    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
513     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
514    
515  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
516    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
517     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
518       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
519    
520  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
521    
# Line 545  ESC_Z to detect the types that may be re Line 603  ESC_Z to detect the types that may be re
603  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
604  character, that code will have to change. */  character, that code will have to change. */
605    
606  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,
607         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,
608         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 };
609    
610    
611  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  /* Opcode table: 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  
612  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.
613    
 To keep stored, compiled patterns compatible, new opcodes should be added  
 immediately before OP_BRA, where (since release 7.0) a gap is left for this  
 purpose.  
   
614  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
615  that follow must also be updated to match. There is also a table called  that follow must also be updated to match. There is also a table called
616  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" in pcre_dfa_exec.c that must be updated. */
# Line 569  enum { Line 622  enum {
622    
623    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
624    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
625    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
626    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
627    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
628    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
629    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
630    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
631    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
632    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
633    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
634    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character */
635    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */
636    OP_PROP,           /* 14 \p (Unicode property) */    OP_NOTPROP,        /* 14 \P (not Unicode property) */
637    OP_ANYNL,          /* 15 \R (any newline sequence) */    OP_PROP,           /* 15 \p (Unicode property) */
638    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */    OP_ANYNL,          /* 16 \R (any newline sequence) */
639    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */
640    OP_EOD,            /* 18 End of data: \z */    OP_HSPACE,         /* 18 \h (horizontal whitespace) */
641      OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */
642    OP_OPT,            /* 19 Set runtime options */    OP_VSPACE,         /* 20 \v (vertical whitespace) */
643    OP_CIRC,           /* 20 Start of line - varies with multiline switch */    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */
644    OP_DOLL,           /* 21 End of line - varies with multiline switch */    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */
645    OP_CHAR,           /* 22 Match one character, casefully */    OP_EOD,            /* 23 End of data: \z */
646    OP_CHARNC,         /* 23 Match one character, caselessly */  
647    OP_NOT,            /* 24 Match one character, not the following one */    OP_OPT,            /* 24 Set runtime options */
648      OP_CIRC,           /* 25 Start of line - varies with multiline switch */
649    OP_STAR,           /* 25 The maximizing and minimizing versions of */    OP_DOLL,           /* 26 End of line - varies with multiline switch */
650    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */    OP_CHAR,           /* 27 Match one character, casefully */
651    OP_PLUS,           /* 27 the minimizing one second. */    OP_CHARNC,         /* 28 Match one character, caselessly */
652    OP_MINPLUS,        /* 28 This first set applies to single characters.*/    OP_NOT,            /* 29 Match one character, not the following one */
653    OP_QUERY,          /* 29 */  
654    OP_MINQUERY,       /* 30 */    OP_STAR,           /* 30 The maximizing and minimizing versions of */
655      OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */
656    OP_UPTO,           /* 31 From 0 to n matches */    OP_PLUS,           /* 32 the minimizing one second. */
657    OP_MINUPTO,        /* 32 */    OP_MINPLUS,        /* 33 This first set applies to single characters.*/
658    OP_EXACT,          /* 33 Exactly n matches */    OP_QUERY,          /* 34 */
659      OP_MINQUERY,       /* 35 */
660    OP_POSSTAR,        /* 34 Possessified star */  
661    OP_POSPLUS,        /* 35 Possessified plus */    OP_UPTO,           /* 36 From 0 to n matches */
662    OP_POSQUERY,       /* 36 Posesssified query */    OP_MINUPTO,        /* 37 */
663    OP_POSUPTO,        /* 37 Possessified upto */    OP_EXACT,          /* 38 Exactly n matches */
664    
665    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */    OP_POSSTAR,        /* 39 Possessified star */
666    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */    OP_POSPLUS,        /* 40 Possessified plus */
667    OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */    OP_POSQUERY,       /* 41 Posesssified query */
668    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */    OP_POSUPTO,        /* 42 Possessified upto */
669    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */  
670    OP_NOTMINQUERY,    /* 43 */    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */
671      OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */
672    OP_NOTUPTO,        /* 44 From 0 to n matches */    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */
673    OP_NOTMINUPTO,     /* 45 */    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */
674    OP_NOTEXACT,       /* 46 Exactly n matches */    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */
675      OP_NOTMINQUERY,    /* 48 */
676    OP_NOTPOSSTAR,     /* 47 Possessified versions */  
677    OP_NOTPOSPLUS,     /* 48 */    OP_NOTUPTO,        /* 49 From 0 to n matches */
678    OP_NOTPOSQUERY,    /* 49 */    OP_NOTMINUPTO,     /* 50 */
679    OP_NOTPOSUPTO,     /* 50 */    OP_NOTEXACT,       /* 51 Exactly n matches */
680    
681    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */    OP_NOTPOSSTAR,     /* 52 Possessified versions */
682    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */    OP_NOTPOSPLUS,     /* 53 */
683    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */    OP_NOTPOSQUERY,    /* 54 */
684    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */    OP_NOTPOSUPTO,     /* 55 */
685    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */  
686    OP_TYPEMINQUERY,   /* 56 */    OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */
687      OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */
688    OP_TYPEUPTO,       /* 57 From 0 to n matches */    OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */
689    OP_TYPEMINUPTO,    /* 58 */    OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */
690    OP_TYPEEXACT,      /* 59 Exactly n matches */    OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */
691      OP_TYPEMINQUERY,   /* 61 */
692    OP_TYPEPOSSTAR,    /* 60 Possessified versions */  
693    OP_TYPEPOSPLUS,    /* 61 */    OP_TYPEUPTO,       /* 62 From 0 to n matches */
694    OP_TYPEPOSQUERY,   /* 62 */    OP_TYPEMINUPTO,    /* 63 */
695    OP_TYPEPOSUPTO,    /* 63 */    OP_TYPEEXACT,      /* 64 Exactly n matches */
696    
697    OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */    OP_TYPEPOSSTAR,    /* 65 Possessified versions */
698    OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */    OP_TYPEPOSPLUS,    /* 66 */
699    OP_CRPLUS,         /* 66 the minimizing one second. These codes must */    OP_TYPEPOSQUERY,   /* 67 */
700    OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */    OP_TYPEPOSUPTO,    /* 68 */
701    OP_CRQUERY,        /* 68 These are for character classes and back refs */  
702    OP_CRMINQUERY,     /* 69 */    OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */
703    OP_CRRANGE,        /* 70 These are different to the three sets above. */    OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */
704    OP_CRMINRANGE,     /* 71 */    OP_CRPLUS,         /* 71 the minimizing one second. These codes must */
705      OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */
706      OP_CRQUERY,        /* 73 These are for character classes and back refs */
707      OP_CRMINQUERY,     /* 74 */
708      OP_CRRANGE,        /* 75 These are different to the three sets above. */
709      OP_CRMINRANGE,     /* 76 */
710    
711    OP_CLASS,          /* 72 Match a character class, chars < 256 only */    OP_CLASS,          /* 77 Match a character class, chars < 256 only */
712    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative
713                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
714                             character > 255 is encountered. */                             character > 255 is encountered. */
715    
716    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the
717                             class. This does both positive and negative. */                             class. This does both positive and negative. */
718    
719    OP_REF,            /* 75 Match a back reference */    OP_REF,            /* 80 Match a back reference */
720    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */
721    OP_CALLOUT,        /* 77 Call out to external function if provided */    OP_CALLOUT,        /* 82 Call out to external function if provided */
722    
723    OP_ALT,            /* 78 Start of alternation */    OP_ALT,            /* 83 Start of alternation */
724    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 */
725    OP_KETRMAX,        /* 80 These two must remain together and in this */    OP_KETRMAX,        /* 85 These two must remain together and in this */
726    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */
727    
728    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
729    
730    OP_ASSERT,         /* 82 Positive lookahead */    OP_ASSERT,         /* 87 Positive lookahead */
731    OP_ASSERT_NOT,     /* 83 Negative lookahead */    OP_ASSERT_NOT,     /* 88 Negative lookahead */
732    OP_ASSERTBACK,     /* 84 Positive lookbehind */    OP_ASSERTBACK,     /* 89 Positive lookbehind */
733    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */
734    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */
735    
736    /* 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,
737    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. */
738    
739    OP_ONCE,           /* 87 Atomic group */    OP_ONCE,           /* 92 Atomic group */
740    OP_BRA,            /* 88 Start of non-capturing bracket */    OP_BRA,            /* 93 Start of non-capturing bracket */
741    OP_CBRA,           /* 89 Start of capturing bracket */    OP_CBRA,           /* 94 Start of capturing bracket */
742    OP_COND,           /* 90 Conditional group */    OP_COND,           /* 95 Conditional group */
743    
744    /* 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
745    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
746    
747    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */
748    OP_SCBRA,          /* 92 Start of capturing bracket, check empty */    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */
749    OP_SCOND,          /* 93 Conditional group, check empty */    OP_SCOND,          /* 98 Conditional group, check empty */
750    
751    OP_CREF,           /* 94 Used to hold a capture number as condition */    OP_CREF,           /* 99 Used to hold a capture number as condition */
752    OP_RREF,           /* 95 Used to hold a recursion number as condition */    OP_RREF,           /* 100 Used to hold a recursion number as condition */
753    OP_DEF,            /* 96 The DEFINE condition */    OP_DEF,            /* 101 The DEFINE condition */
754    
755    OP_BRAZERO,        /* 97 These two must remain together and in this */    OP_BRAZERO,        /* 102 These two must remain together and in this */
756    OP_BRAMINZERO      /* 98 order. */    OP_BRAMINZERO,     /* 103 order. */
757    
758      /* These are backtracking control verbs */
759    
760      OP_PRUNE,          /* 104 */
761      OP_SKIP,           /* 105 */
762      OP_THEN,           /* 106 */
763      OP_COMMIT,         /* 107 */
764    
765      /* These are forced failure and success verbs */
766    
767      OP_FAIL,           /* 108 */
768      OP_ACCEPT          /* 109 */
769  };  };
770    
771    
# Line 703  enum { Line 773  enum {
773  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
774    
775  #define OP_NAME_LIST \  #define OP_NAME_LIST \
776    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
777    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \
778    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
779    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
780    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
781    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
782    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 718  for debugging. The macro is referenced o Line 788  for debugging. The macro is referenced o
788    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
789    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
790    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
791    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
792    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
793      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT"
794    
795    
796  /* 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 733  in UTF-8 mode. The code that uses this t Line 804  in UTF-8 mode. The code that uses this t
804    
805  #define OP_LENGTHS \  #define OP_LENGTHS \
806    1,                             /* End                                    */ \    1,                             /* End                                    */ \
807    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                     */ \
808      1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
809    1, 1,                          /* Any, Anybyte                           */ \    1, 1,                          /* Any, Anybyte                           */ \
810    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
811      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
812    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
813    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
814    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 781  in UTF-8 mode. The code that uses this t Line 854  in UTF-8 mode. The code that uses this t
854    3,                             /* RREF                                   */ \    3,                             /* RREF                                   */ \
855    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
856    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
857      1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
858      1, 1                           /* FAIL, ACCEPT                           */
859    
860    
861  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
# Line 795  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 870  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
870         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
871         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
872         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
873         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
874           ERR60, ERR61, ERR62 };
875    
876  /* 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
877  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 817  NOTE NOTE NOTE: Line 893  NOTE NOTE NOTE:
893  typedef struct real_pcre {  typedef struct real_pcre {
894    pcre_uint32 magic_number;    pcre_uint32 magic_number;
895    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
896    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
897    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
898      pcre_uint16 dummy1;             /* For future use */
899    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
900    pcre_uint16 top_backref;    pcre_uint16 top_backref;
901    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 862  typedef struct compile_data { Line 938  typedef struct compile_data {
938    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
939    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
940    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
941      int  external_flags;          /* External flag bits to be set */
942    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
943    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
944    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
945    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
946    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 884  typedef struct recursion_info { Line 961  typedef struct recursion_info {
961    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
962    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
963    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 */
964    USPTR save_start;             /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
965    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
966    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
967  } recursion_info;  } recursion_info;
968    
 /* 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;  
   
969  /* 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
970  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
971  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops. */
# Line 932  typedef struct match_data { Line 999  typedef struct match_data {
999    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1000    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1001    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1002      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1003    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1004    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1005    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1006    USPTR  start_match;           /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1007    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1008    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1009    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
# Line 944  typedef struct match_data { Line 1012  typedef struct match_data {
1012    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1013    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1014    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 */  
1015  } match_data;  } match_data;
1016    
1017  /* 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 997  total length. */ Line 1064  total length. */
1064  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1065    
1066  /* Layout of the UCP type table that translates property names into types and  /* Layout of the UCP type table that translates property names into types and
1067  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1068    relocations in shared libraries, it now has an offset into a single string
1069    instead. */
1070    
1071  typedef struct {  typedef struct {
1072    const char *name;    pcre_uint16 name_offset;
1073    pcre_uint16 type;    pcre_uint16 type;
1074    pcre_uint16 value;    pcre_uint16 value;
1075  } ucp_type_table;  } ucp_type_table;
# Line 1018  extern const uschar _pcre_utf8_table4[]; Line 1087  extern const uschar _pcre_utf8_table4[];
1087    
1088  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1089    
1090    extern const char   _pcre_utt_names[];
1091  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1092  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1093    
# Line 1030  extern const uschar _pcre_OP_lengths[]; Line 1100  extern const uschar _pcre_OP_lengths[];
1100  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
1101  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1102    
1103  extern BOOL         _pcre_is_newline(const uschar *, const uschar *, int *,  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
1104                        BOOL);                        int *, BOOL);
1105  extern int          _pcre_ord2utf8(int, uschar *);  extern int          _pcre_ord2utf8(int, uschar *);
1106  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1107                        const pcre_study_data *, pcre_study_data *);                        const pcre_study_data *, pcre_study_data *);
1108  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);
1109  extern unsigned int _pcre_ucp_othercase(const unsigned int);  extern unsigned int _pcre_ucp_othercase(const unsigned int);
1110  extern int          _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1111  extern BOOL         _pcre_was_newline(const uschar *, const uschar *, int *,  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1112                        BOOL);                        int *, BOOL);
1113  extern BOOL         _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_xclass(int, const uschar *);
1114    
1115  #endif  #endif
Legend:
Removed from v.108  
changed lines
  Added in v.268
  ViewVC Help
Powered by ViewVC 1.1.5