/[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 168 by ph10, Tue May 29 15:18:18 2007 UTC revision 345 by ph10, Mon Apr 28 15:10:02 2008 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) 1997-2007 University of Cambridge             Copyright (c) 1997-2008 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 112  PCRE_EXP_DATA_DEFN only if they are not Line 108  PCRE_EXP_DATA_DEFN only if they are not
108    
109  #ifndef PCRE_EXP_DECL  #ifndef PCRE_EXP_DECL
110  #  ifdef _WIN32  #  ifdef _WIN32
111  #    ifdef DLL_EXPORT  #    ifndef PCRE_STATIC
112  #      define PCRE_EXP_DECL       extern __declspec(dllexport)  #      define PCRE_EXP_DECL       extern __declspec(dllexport)
113  #      define PCRE_EXP_DEFN       __declspec(dllexport)  #      define PCRE_EXP_DEFN       __declspec(dllexport)
114  #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)  #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
# Line 121  PCRE_EXP_DATA_DEFN only if they are not Line 117  PCRE_EXP_DATA_DEFN only if they are not
117  #      define PCRE_EXP_DEFN  #      define PCRE_EXP_DEFN
118  #      define PCRE_EXP_DATA_DEFN  #      define PCRE_EXP_DATA_DEFN
119  #    endif  #    endif
 #  
120  #  else  #  else
121  #    ifdef __cplusplus  #    ifdef __cplusplus
122  #      define PCRE_EXP_DECL       extern "C"  #      define PCRE_EXP_DECL       extern "C"
# Line 234  must begin with PCRE_. */ Line 229  must begin with PCRE_. */
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    
232  #include <pcre.h>  #include "pcre.h"
233  #include "ucp.h"  #include "ucp.h"
234    
235  /* When compiling for use with the Virtual Pascal compiler, these functions  /* When compiling for use with the Virtual Pascal compiler, these functions
# Line 363  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 NEXTCHAR(p) p++;
367  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
368  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
369  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
370  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
371  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
372  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
373    
374  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
375    
376    /* Advance a character pointer one byte in non-UTF-8 mode and by one character
377    in UTF-8 mode. */
378    
379    #define NEXTCHAR(p) \
380      p++; \
381      if (utf8) { while((*p & 0xc0) == 0x80) p++; }
382    
383  /* Get the next UTF-8 character, not advancing the pointer. This is called when  /* Get the next UTF-8 character, not advancing the pointer. This is called when
384  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
385    
# Line 464  if there are extra bytes. This is called Line 469  if there are extra bytes. This is called
469      }      }
470    
471  /* 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
472  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
473    because almost all calls are already within a block of UTF-8 only code. */
474    
475  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
476    
477  #endif  #endif
478    
# Line 483  Standard C system should have one. */ Line 489  Standard C system should have one. */
489    
490  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
491    
492  /* Private options flags start at the most significant end of the four bytes.  /* Private flags containing information about the compiled regex. They used to
493  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
494  sure they don't overlap! The bits are getting a bit scarce now -- when we run  are in a 16-bit flags word. */
495  out, there is a dummy word in the structure that could be used for the private  
496  bits. */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
497    #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
498  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
499  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
500  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
501  #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 */  
502    
503  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
504    
# Line 509  time, run time, or study time, respectiv Line 514  time, run time, or study time, respectiv
514    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
515     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
516     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| \
517     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
518       PCRE_JAVASCRIPT_COMPAT)
519    
520  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
521    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
522     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
523    
524  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
525    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
526     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
527       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
528    
529  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
530    
# Line 598  contain UTF-8 characters with values gre Line 605  contain UTF-8 characters with values gre
605  value such as \n. They must have non-zero values, as check_escape() returns  value such as \n. They must have non-zero values, as check_escape() returns
606  their negation. Also, they must appear in the same order as in the opcode  their negation. Also, they must appear in the same order as in the opcode
607  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  definitions below, up to ESC_z. There's a dummy for OP_ANY because it
608  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
609  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  (which is used for [^] in JavaScript compatibility mode).
 There are two tests in the code for an escape greater than ESC_b and less than  
 ESC_Z to detect the types that may be repeated. These are the types that  
 consume characters. If any new escapes are put in between that don't consume a  
 character, that code will have to change. */  
   
 enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,  
        ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,  
        ESC_E, ESC_Q, ESC_k, ESC_REF };  
610    
611    The final escape must be ESC_REF as subsequent values are used for
612    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
613    greater than ESC_b and less than ESC_Z to detect the types that may be
614    repeated. These are the types that consume characters. If any new escapes are
615    put in between that don't consume a character, that code will have to change.
616    */
617    
618    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
619           ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
620           ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
621           ESC_REF };
622    
 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  
 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  
 OP_EOD must correspond in order to the list of escapes immediately above.  
623    
624  To keep stored, compiled patterns compatible, new opcodes should be added  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
625  immediately before OP_BRA, where (since release 7.0) a gap is left for this  OP_EOD must correspond in order to the list of escapes immediately above.
 purpose.  
626    
627  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
628  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
# Line 629  enum { Line 635  enum {
635    
636    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
637    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
638    OP_SET_SOM,        /* 3 Set start of match (\K) */    OP_SET_SOM,        /* 3 Set start of match (\K) */
639    OP_NOT_WORD_BOUNDARY,  /*  4 \B */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
640    OP_WORD_BOUNDARY,      /*  5 \b */    OP_WORD_BOUNDARY,      /*  5 \b */
641    OP_NOT_DIGIT,          /*  6 \D */    OP_NOT_DIGIT,          /*  6 \D */
# Line 638  enum { Line 644  enum {
644    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
645    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
646    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
647    OP_ANY,            /* 12 Match any character */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
648    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
649    OP_NOTPROP,        /* 14 \P (not Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
650    OP_PROP,           /* 15 \p (Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
651    OP_ANYNL,          /* 16 \R (any newline sequence) */    OP_PROP,           /* 16 \p (Unicode property) */
652    OP_EXTUNI,         /* 17 \X (extended Unicode sequence */    OP_ANYNL,          /* 17 \R (any newline sequence) */
653    OP_EODN,           /* 18 End of data or \n at end of data: \Z. */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
654    OP_EOD,            /* 19 End of data: \z */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
655      OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
656    OP_OPT,            /* 20 Set runtime options */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
657    OP_CIRC,           /* 21 Start of line - varies with multiline switch */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
658    OP_DOLL,           /* 22 End of line - varies with multiline switch */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
659    OP_CHAR,           /* 23 Match one character, casefully */    OP_EOD,            /* 24 End of data: \z */
660    OP_CHARNC,         /* 24 Match one character, caselessly */  
661    OP_NOT,            /* 25 Match one character, not the following one */    OP_OPT,            /* 25 Set runtime options */
662      OP_CIRC,           /* 26 Start of line - varies with multiline switch */
663    OP_STAR,           /* 26 The maximizing and minimizing versions of */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
664    OP_MINSTAR,        /* 27 these six opcodes must come in pairs, with */    OP_CHAR,           /* 28 Match one character, casefully */
665    OP_PLUS,           /* 28 the minimizing one second. */    OP_CHARNC,         /* 29 Match one character, caselessly */
666    OP_MINPLUS,        /* 29 This first set applies to single characters.*/    OP_NOT,            /* 30 Match one character, not the following one */
667    OP_QUERY,          /* 30 */  
668    OP_MINQUERY,       /* 31 */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
669      OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
670    OP_UPTO,           /* 32 From 0 to n matches */    OP_PLUS,           /* 33 the minimizing one second. */
671    OP_MINUPTO,        /* 33 */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
672    OP_EXACT,          /* 34 Exactly n matches */    OP_QUERY,          /* 35 */
673      OP_MINQUERY,       /* 36 */
674    OP_POSSTAR,        /* 35 Possessified star */  
675    OP_POSPLUS,        /* 36 Possessified plus */    OP_UPTO,           /* 37 From 0 to n matches */
676    OP_POSQUERY,       /* 37 Posesssified query */    OP_MINUPTO,        /* 38 */
677    OP_POSUPTO,        /* 38 Possessified upto */    OP_EXACT,          /* 39 Exactly n matches */
678    
679    OP_NOTSTAR,        /* 39 The maximizing and minimizing versions of */    OP_POSSTAR,        /* 40 Possessified star */
680    OP_NOTMINSTAR,     /* 40 these six opcodes must come in pairs, with */    OP_POSPLUS,        /* 41 Possessified plus */
681    OP_NOTPLUS,        /* 41 the minimizing one second. They must be in */    OP_POSQUERY,       /* 42 Posesssified query */
682    OP_NOTMINPLUS,     /* 42 exactly the same order as those above. */    OP_POSUPTO,        /* 43 Possessified upto */
683    OP_NOTQUERY,       /* 43 This set applies to "not" single characters. */  
684    OP_NOTMINQUERY,    /* 44 */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
685      OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
686    OP_NOTUPTO,        /* 45 From 0 to n matches */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
687    OP_NOTMINUPTO,     /* 46 */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
688    OP_NOTEXACT,       /* 47 Exactly n matches */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
689      OP_NOTMINQUERY,    /* 49 */
690    OP_NOTPOSSTAR,     /* 48 Possessified versions */  
691    OP_NOTPOSPLUS,     /* 49 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
692    OP_NOTPOSQUERY,    /* 50 */    OP_NOTMINUPTO,     /* 51 */
693    OP_NOTPOSUPTO,     /* 51 */    OP_NOTEXACT,       /* 52 Exactly n matches */
694    
695    OP_TYPESTAR,       /* 52 The maximizing and minimizing versions of */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
696    OP_TYPEMINSTAR,    /* 53 these six opcodes must come in pairs, with */    OP_NOTPOSPLUS,     /* 54 */
697    OP_TYPEPLUS,       /* 54 the minimizing one second. These codes must */    OP_NOTPOSQUERY,    /* 55 */
698    OP_TYPEMINPLUS,    /* 55 be in exactly the same order as those above. */    OP_NOTPOSUPTO,     /* 56 */
699    OP_TYPEQUERY,      /* 56 This set applies to character types such as \d */  
700    OP_TYPEMINQUERY,   /* 57 */    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
701      OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
702    OP_TYPEUPTO,       /* 58 From 0 to n matches */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
703    OP_TYPEMINUPTO,    /* 59 */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
704    OP_TYPEEXACT,      /* 60 Exactly n matches */    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
705      OP_TYPEMINQUERY,   /* 62 */
706    OP_TYPEPOSSTAR,    /* 61 Possessified versions */  
707    OP_TYPEPOSPLUS,    /* 62 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
708    OP_TYPEPOSQUERY,   /* 63 */    OP_TYPEMINUPTO,    /* 64 */
709    OP_TYPEPOSUPTO,    /* 64 */    OP_TYPEEXACT,      /* 65 Exactly n matches */
710    
711    OP_CRSTAR,         /* 65 The maximizing and minimizing versions of */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
712    OP_CRMINSTAR,      /* 66 all these opcodes must come in pairs, with */    OP_TYPEPOSPLUS,    /* 67 */
713    OP_CRPLUS,         /* 67 the minimizing one second. These codes must */    OP_TYPEPOSQUERY,   /* 68 */
714    OP_CRMINPLUS,      /* 68 be in exactly the same order as those above. */    OP_TYPEPOSUPTO,    /* 69 */
715    OP_CRQUERY,        /* 69 These are for character classes and back refs */  
716    OP_CRMINQUERY,     /* 70 */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
717    OP_CRRANGE,        /* 71 These are different to the three sets above. */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
718    OP_CRMINRANGE,     /* 72 */    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
719      OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
720      OP_CRQUERY,        /* 74 These are for character classes and back refs */
721      OP_CRMINQUERY,     /* 75 */
722      OP_CRRANGE,        /* 76 These are different to the three sets above. */
723      OP_CRMINRANGE,     /* 77 */
724    
725    OP_CLASS,          /* 73 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
726    OP_NCLASS,         /* 74 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
727                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
728                             character > 255 is encountered. */                             character > 255 is encountered. */
729    
730    OP_XCLASS,         /* 75 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
731                             class. This does both positive and negative. */                             class. This does both positive and negative. */
732    
733    OP_REF,            /* 76 Match a back reference */    OP_REF,            /* 81 Match a back reference */
734    OP_RECURSE,        /* 77 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
735    OP_CALLOUT,        /* 78 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
736    
737    OP_ALT,            /* 79 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
738    OP_KET,            /* 80 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */
739    OP_KETRMAX,        /* 81 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
740    OP_KETRMIN,        /* 82 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
741    
742    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
743    
744    OP_ASSERT,         /* 83 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
745    OP_ASSERT_NOT,     /* 84 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
746    OP_ASSERTBACK,     /* 85 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
747    OP_ASSERTBACK_NOT, /* 86 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
748    OP_REVERSE,        /* 87 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
749    
750    /* 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,
751    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. */
752    
753    OP_ONCE,           /* 88 Atomic group */    OP_ONCE,           /* 93 Atomic group */
754    OP_BRA,            /* 89 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
755    OP_CBRA,           /* 90 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
756    OP_COND,           /* 91 Conditional group */    OP_COND,           /* 96 Conditional group */
757    
758    /* 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
759    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
760    
761    OP_SBRA,           /* 92 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
762    OP_SCBRA,          /* 93 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
763    OP_SCOND,          /* 94 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
764    
765    OP_CREF,           /* 95 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
766    OP_RREF,           /* 96 Used to hold a recursion number as condition */    OP_RREF,           /* 101 Used to hold a recursion number as condition */
767    OP_DEF,            /* 97 The DEFINE condition */    OP_DEF,            /* 102 The DEFINE condition */
768    
769    OP_BRAZERO,        /* 98 These two must remain together and in this */    OP_BRAZERO,        /* 103 These two must remain together and in this */
770    OP_BRAMINZERO      /* 99 order. */    OP_BRAMINZERO,     /* 104 order. */
771    
772      /* These are backtracking control verbs */
773    
774      OP_PRUNE,          /* 105 */
775      OP_SKIP,           /* 106 */
776      OP_THEN,           /* 107 */
777      OP_COMMIT,         /* 108 */
778    
779      /* These are forced failure and success verbs */
780    
781      OP_FAIL,           /* 109 */
782      OP_ACCEPT,         /* 110 */
783    
784      /* This is used to skip a subpattern with a {0} quantifier */
785    
786      OP_SKIPZERO        /* 111 */
787  };  };
788    
789    
# Line 765  for debugging. The macro is referenced o Line 792  for debugging. The macro is referenced o
792    
793  #define OP_NAME_LIST \  #define OP_NAME_LIST \
794    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
795    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
796    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
797    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
798    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
799    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
800    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 779  for debugging. The macro is referenced o Line 806  for debugging. The macro is referenced o
806    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
807    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
808    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
809    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
810    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
811      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
812      "Skip zero"
813    
814    
815  /* 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 796  in UTF-8 mode. The code that uses this t Line 825  in UTF-8 mode. The code that uses this t
825    1,                             /* End                                    */ \    1,                             /* End                                    */ \
826    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
827    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
828    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
829    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
830      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
831    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
832    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
833    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 843  in UTF-8 mode. The code that uses this t Line 873  in UTF-8 mode. The code that uses this t
873    3,                             /* RREF                                   */ \    3,                             /* RREF                                   */ \
874    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
875    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
876      1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
877      1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
878    
879    
880  /* 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 857  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 889  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
889         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
890         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
891         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
892         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
893           ERR60, ERR61, ERR62, ERR63, ERR64 };
894    
895  /* 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
896  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 879  NOTE NOTE NOTE: Line 912  NOTE NOTE NOTE:
912  typedef struct real_pcre {  typedef struct real_pcre {
913    pcre_uint32 magic_number;    pcre_uint32 magic_number;
914    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
915    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
916    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
917      pcre_uint16 dummy1;             /* For future use */
918    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
919    pcre_uint16 top_backref;    pcre_uint16 top_backref;
920    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 920  typedef struct compile_data { Line 953  typedef struct compile_data {
953    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
954    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
955    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
956    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
957      int  final_bracount;          /* Saved value after first pass */
958    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
959    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
960    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
961      int  external_flags;          /* External flag bits to be set */
962    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
963    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
964    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
965    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
966    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 980  typedef struct match_data { Line 1015  typedef struct match_data {
1015    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1016    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1017    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1018      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1019    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1020    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1021    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1022    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1023      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1024    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1025    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1026    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
# Line 1020  typedef struct dfa_match_data { Line 1057  typedef struct dfa_match_data {
1057  #define ctype_letter  0x02  #define ctype_letter  0x02
1058  #define ctype_digit   0x04  #define ctype_digit   0x04
1059  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1060  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1061  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1062    
1063  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
# Line 1048  total length. */ Line 1085  total length. */
1085  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1086    
1087  /* 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
1088  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1089    relocations in shared libraries, it now has an offset into a single string
1090    instead. */
1091    
1092  typedef struct {  typedef struct {
1093    const char *name;    pcre_uint16 name_offset;
1094    pcre_uint16 type;    pcre_uint16 type;
1095    pcre_uint16 value;    pcre_uint16 value;
1096  } ucp_type_table;  } ucp_type_table;
# Line 1069  extern const uschar _pcre_utf8_table4[]; Line 1108  extern const uschar _pcre_utf8_table4[];
1108    
1109  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1110    
1111    extern const char   _pcre_utt_names[];
1112  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1113  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1114    

Legend:
Removed from v.168  
changed lines
  Added in v.345

  ViewVC Help
Powered by ViewVC 1.1.5