/[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 336 by ph10, Sat Apr 12 15:59:03 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 605  ESC_Z to detect the types that may be re Line 612  ESC_Z to detect the types that may be re
612  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
613  character, that code will have to change. */  character, that code will have to change. */
614    
615  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
616         ESC_W, 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,
617         ESC_E, ESC_Q, ESC_k, ESC_REF };         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k, ESC_REF };
618    
619    
620  /* 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  
621  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.
622    
 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.  
   
623  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
624  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
625  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" in pcre_dfa_exec.c that must be updated. */
# Line 629  enum { Line 631  enum {
631    
632    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
633    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
634    OP_SET_SOM,        /* 3 Set start of match (\K) */    OP_SET_SOM,        /* 3 Set start of match (\K) */
635    OP_NOT_WORD_BOUNDARY,  /*  4 \B */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
636    OP_WORD_BOUNDARY,      /*  5 \b */    OP_WORD_BOUNDARY,      /*  5 \b */
637    OP_NOT_DIGIT,          /*  6 \D */    OP_NOT_DIGIT,          /*  6 \D */
# Line 643  enum { Line 645  enum {
645    OP_NOTPROP,        /* 14 \P (not Unicode property) */    OP_NOTPROP,        /* 14 \P (not Unicode property) */
646    OP_PROP,           /* 15 \p (Unicode property) */    OP_PROP,           /* 15 \p (Unicode property) */
647    OP_ANYNL,          /* 16 \R (any newline sequence) */    OP_ANYNL,          /* 16 \R (any newline sequence) */
648    OP_EXTUNI,         /* 17 \X (extended Unicode sequence */    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */
649    OP_EODN,           /* 18 End of data or \n at end of data: \Z. */    OP_HSPACE,         /* 18 \h (horizontal whitespace) */
650    OP_EOD,            /* 19 End of data: \z */    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */
651      OP_VSPACE,         /* 20 \v (vertical whitespace) */
652    OP_OPT,            /* 20 Set runtime options */    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */
653    OP_CIRC,           /* 21 Start of line - varies with multiline switch */    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */
654    OP_DOLL,           /* 22 End of line - varies with multiline switch */    OP_EOD,            /* 23 End of data: \z */
655    OP_CHAR,           /* 23 Match one character, casefully */  
656    OP_CHARNC,         /* 24 Match one character, caselessly */    OP_OPT,            /* 24 Set runtime options */
657    OP_NOT,            /* 25 Match one character, not the following one */    OP_CIRC,           /* 25 Start of line - varies with multiline switch */
658      OP_DOLL,           /* 26 End of line - varies with multiline switch */
659    OP_STAR,           /* 26 The maximizing and minimizing versions of */    OP_CHAR,           /* 27 Match one character, casefully */
660    OP_MINSTAR,        /* 27 these six opcodes must come in pairs, with */    OP_CHARNC,         /* 28 Match one character, caselessly */
661    OP_PLUS,           /* 28 the minimizing one second. */    OP_NOT,            /* 29 Match one character, not the following one */
662    OP_MINPLUS,        /* 29 This first set applies to single characters.*/  
663    OP_QUERY,          /* 30 */    OP_STAR,           /* 30 The maximizing and minimizing versions of */
664    OP_MINQUERY,       /* 31 */    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */
665      OP_PLUS,           /* 32 the minimizing one second. */
666    OP_UPTO,           /* 32 From 0 to n matches */    OP_MINPLUS,        /* 33 This first set applies to single characters.*/
667    OP_MINUPTO,        /* 33 */    OP_QUERY,          /* 34 */
668    OP_EXACT,          /* 34 Exactly n matches */    OP_MINQUERY,       /* 35 */
669    
670    OP_POSSTAR,        /* 35 Possessified star */    OP_UPTO,           /* 36 From 0 to n matches */
671    OP_POSPLUS,        /* 36 Possessified plus */    OP_MINUPTO,        /* 37 */
672    OP_POSQUERY,       /* 37 Posesssified query */    OP_EXACT,          /* 38 Exactly n matches */
673    OP_POSUPTO,        /* 38 Possessified upto */  
674      OP_POSSTAR,        /* 39 Possessified star */
675    OP_NOTSTAR,        /* 39 The maximizing and minimizing versions of */    OP_POSPLUS,        /* 40 Possessified plus */
676    OP_NOTMINSTAR,     /* 40 these six opcodes must come in pairs, with */    OP_POSQUERY,       /* 41 Posesssified query */
677    OP_NOTPLUS,        /* 41 the minimizing one second. They must be in */    OP_POSUPTO,        /* 42 Possessified upto */
678    OP_NOTMINPLUS,     /* 42 exactly the same order as those above. */  
679    OP_NOTQUERY,       /* 43 This set applies to "not" single characters. */    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */
680    OP_NOTMINQUERY,    /* 44 */    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */
681      OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */
682    OP_NOTUPTO,        /* 45 From 0 to n matches */    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */
683    OP_NOTMINUPTO,     /* 46 */    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */
684    OP_NOTEXACT,       /* 47 Exactly n matches */    OP_NOTMINQUERY,    /* 48 */
685    
686    OP_NOTPOSSTAR,     /* 48 Possessified versions */    OP_NOTUPTO,        /* 49 From 0 to n matches */
687    OP_NOTPOSPLUS,     /* 49 */    OP_NOTMINUPTO,     /* 50 */
688    OP_NOTPOSQUERY,    /* 50 */    OP_NOTEXACT,       /* 51 Exactly n matches */
689    OP_NOTPOSUPTO,     /* 51 */  
690      OP_NOTPOSSTAR,     /* 52 Possessified versions */
691    OP_TYPESTAR,       /* 52 The maximizing and minimizing versions of */    OP_NOTPOSPLUS,     /* 53 */
692    OP_TYPEMINSTAR,    /* 53 these six opcodes must come in pairs, with */    OP_NOTPOSQUERY,    /* 54 */
693    OP_TYPEPLUS,       /* 54 the minimizing one second. These codes must */    OP_NOTPOSUPTO,     /* 55 */
694    OP_TYPEMINPLUS,    /* 55 be in exactly the same order as those above. */  
695    OP_TYPEQUERY,      /* 56 This set applies to character types such as \d */    OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */
696    OP_TYPEMINQUERY,   /* 57 */    OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */
697      OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */
698    OP_TYPEUPTO,       /* 58 From 0 to n matches */    OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */
699    OP_TYPEMINUPTO,    /* 59 */    OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */
700    OP_TYPEEXACT,      /* 60 Exactly n matches */    OP_TYPEMINQUERY,   /* 61 */
701    
702    OP_TYPEPOSSTAR,    /* 61 Possessified versions */    OP_TYPEUPTO,       /* 62 From 0 to n matches */
703    OP_TYPEPOSPLUS,    /* 62 */    OP_TYPEMINUPTO,    /* 63 */
704    OP_TYPEPOSQUERY,   /* 63 */    OP_TYPEEXACT,      /* 64 Exactly n matches */
705    OP_TYPEPOSUPTO,    /* 64 */  
706      OP_TYPEPOSSTAR,    /* 65 Possessified versions */
707    OP_CRSTAR,         /* 65 The maximizing and minimizing versions of */    OP_TYPEPOSPLUS,    /* 66 */
708    OP_CRMINSTAR,      /* 66 all these opcodes must come in pairs, with */    OP_TYPEPOSQUERY,   /* 67 */
709    OP_CRPLUS,         /* 67 the minimizing one second. These codes must */    OP_TYPEPOSUPTO,    /* 68 */
710    OP_CRMINPLUS,      /* 68 be in exactly the same order as those above. */  
711    OP_CRQUERY,        /* 69 These are for character classes and back refs */    OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */
712    OP_CRMINQUERY,     /* 70 */    OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */
713    OP_CRRANGE,        /* 71 These are different to the three sets above. */    OP_CRPLUS,         /* 71 the minimizing one second. These codes must */
714    OP_CRMINRANGE,     /* 72 */    OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */
715      OP_CRQUERY,        /* 73 These are for character classes and back refs */
716      OP_CRMINQUERY,     /* 74 */
717      OP_CRRANGE,        /* 75 These are different to the three sets above. */
718      OP_CRMINRANGE,     /* 76 */
719    
720    OP_CLASS,          /* 73 Match a character class, chars < 256 only */    OP_CLASS,          /* 77 Match a character class, chars < 256 only */
721    OP_NCLASS,         /* 74 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative
722                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
723                             character > 255 is encountered. */                             character > 255 is encountered. */
724    
725    OP_XCLASS,         /* 75 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the
726                             class. This does both positive and negative. */                             class. This does both positive and negative. */
727    
728    OP_REF,            /* 76 Match a back reference */    OP_REF,            /* 80 Match a back reference */
729    OP_RECURSE,        /* 77 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */
730    OP_CALLOUT,        /* 78 Call out to external function if provided */    OP_CALLOUT,        /* 82 Call out to external function if provided */
731    
732    OP_ALT,            /* 79 Start of alternation */    OP_ALT,            /* 83 Start of alternation */
733    OP_KET,            /* 80 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */
734    OP_KETRMAX,        /* 81 These two must remain together and in this */    OP_KETRMAX,        /* 85 These two must remain together and in this */
735    OP_KETRMIN,        /* 82 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */
736    
737    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
738    
739    OP_ASSERT,         /* 83 Positive lookahead */    OP_ASSERT,         /* 87 Positive lookahead */
740    OP_ASSERT_NOT,     /* 84 Negative lookahead */    OP_ASSERT_NOT,     /* 88 Negative lookahead */
741    OP_ASSERTBACK,     /* 85 Positive lookbehind */    OP_ASSERTBACK,     /* 89 Positive lookbehind */
742    OP_ASSERTBACK_NOT, /* 86 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */
743    OP_REVERSE,        /* 87 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */
744    
745    /* 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,
746    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. */
747    
748    OP_ONCE,           /* 88 Atomic group */    OP_ONCE,           /* 92 Atomic group */
749    OP_BRA,            /* 89 Start of non-capturing bracket */    OP_BRA,            /* 93 Start of non-capturing bracket */
750    OP_CBRA,           /* 90 Start of capturing bracket */    OP_CBRA,           /* 94 Start of capturing bracket */
751    OP_COND,           /* 91 Conditional group */    OP_COND,           /* 95 Conditional group */
752    
753    /* 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
754    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
755    
756    OP_SBRA,           /* 92 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */
757    OP_SCBRA,          /* 93 Start of capturing bracket, check empty */    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */
758    OP_SCOND,          /* 94 Conditional group, check empty */    OP_SCOND,          /* 98 Conditional group, check empty */
759    
760    OP_CREF,           /* 95 Used to hold a capture number as condition */    OP_CREF,           /* 99 Used to hold a capture number as condition */
761    OP_RREF,           /* 96 Used to hold a recursion number as condition */    OP_RREF,           /* 100 Used to hold a recursion number as condition */
762    OP_DEF,            /* 97 The DEFINE condition */    OP_DEF,            /* 101 The DEFINE condition */
763    
764    OP_BRAZERO,        /* 98 These two must remain together and in this */    OP_BRAZERO,        /* 102 These two must remain together and in this */
765    OP_BRAMINZERO      /* 99 order. */    OP_BRAMINZERO,     /* 103 order. */
766    
767      /* These are backtracking control verbs */
768    
769      OP_PRUNE,          /* 104 */
770      OP_SKIP,           /* 105 */
771      OP_THEN,           /* 106 */
772      OP_COMMIT,         /* 107 */
773    
774      /* These are forced failure and success verbs */
775    
776      OP_FAIL,           /* 108 */
777      OP_ACCEPT,         /* 109 */
778    
779      /* This is used to skip a subpattern with a {0} quantifier */
780    
781      OP_SKIPZERO        /* 110 */
782  };  };
783    
784    
# Line 766  for debugging. The macro is referenced o Line 788  for debugging. The macro is referenced o
788  #define OP_NAME_LIST \  #define OP_NAME_LIST \
789    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
790    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \
791    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
792    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
793    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
794    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
795    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 779  for debugging. The macro is referenced o Line 801  for debugging. The macro is referenced o
801    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
802    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
803    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
804    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
805    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
806      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
807      "Skip zero"
808    
809    
810  /* 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 797  in UTF-8 mode. The code that uses this t Line 821  in UTF-8 mode. The code that uses this t
821    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
822    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
823    1, 1,                          /* Any, Anybyte                           */ \    1, 1,                          /* Any, Anybyte                           */ \
824    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
825      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
826    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
827    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
828    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 843  in UTF-8 mode. The code that uses this t Line 868  in UTF-8 mode. The code that uses this t
868    3,                             /* RREF                                   */ \    3,                             /* RREF                                   */ \
869    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
870    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
871      1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
872      1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
873    
874    
875  /* 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 884  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
884         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
885         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
886         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
887         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
888           ERR60, ERR61, ERR62, ERR63, ERR64 };
889    
890  /* 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
891  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 907  NOTE NOTE NOTE:
907  typedef struct real_pcre {  typedef struct real_pcre {
908    pcre_uint32 magic_number;    pcre_uint32 magic_number;
909    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
910    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
911    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
912      pcre_uint16 dummy1;             /* For future use */
913    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
914    pcre_uint16 top_backref;    pcre_uint16 top_backref;
915    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 920  typedef struct compile_data { Line 948  typedef struct compile_data {
948    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
949    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
950    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
951    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
952      int  final_bracount;          /* Saved value after first pass */
953    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
954    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
955    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
956      int  external_flags;          /* External flag bits to be set */
957    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
958    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
959    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
960    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
961    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 980  typedef struct match_data { Line 1010  typedef struct match_data {
1010    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1011    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1012    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1013      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1014    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1015    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1016    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1017    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1018      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1019    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1020    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1021    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
# Line 1020  typedef struct dfa_match_data { Line 1052  typedef struct dfa_match_data {
1052  #define ctype_letter  0x02  #define ctype_letter  0x02
1053  #define ctype_digit   0x04  #define ctype_digit   0x04
1054  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1055  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1056  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1057    
1058  /* 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 1080  total length. */
1080  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1081    
1082  /* 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
1083  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1084    relocations in shared libraries, it now has an offset into a single string
1085    instead. */
1086    
1087  typedef struct {  typedef struct {
1088    const char *name;    pcre_uint16 name_offset;
1089    pcre_uint16 type;    pcre_uint16 type;
1090    pcre_uint16 value;    pcre_uint16 value;
1091  } ucp_type_table;  } ucp_type_table;
# Line 1069  extern const uschar _pcre_utf8_table4[]; Line 1103  extern const uschar _pcre_utf8_table4[];
1103    
1104  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1105    
1106    extern const char   _pcre_utt_names[];
1107  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1108  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1109    

Legend:
Removed from v.168  
changed lines
  Added in v.336

  ViewVC Help
Powered by ViewVC 1.1.5