/[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 213 by ph10, Wed Aug 15 11:34:14 2007 UTC revision 349 by ph10, Wed Jul 2 18:42:11 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 140  preprocessor time in standard C environm Line 140  preprocessor time in standard C environm
140    
141  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
142    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
143      typedef short pcre_int16;
144  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
145    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
146      typedef int pcre_int16;
147  #else  #else
148    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
149  #endif  #endif
150    
151  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
152    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
153      typedef int pcre_int32;
154  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
155    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
156      typedef long int pcre_int32;
157  #else  #else
158    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
159  #endif  #endif
# Line 363  never be called in byte mode. To make su Line 367  never be called in byte mode. To make su
367  support is omitted, we don't even define it. */  support is omitted, we don't even define it. */
368    
369  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
370    #define NEXTCHAR(p) p++;
371  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
372  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
373  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
# Line 372  support is omitted, we don't even define Line 377  support is omitted, we don't even define
377    
378  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
379    
380    /* Advance a character pointer one byte in non-UTF-8 mode and by one character
381    in UTF-8 mode. */
382    
383    #define NEXTCHAR(p) \
384      p++; \
385      if (utf8) { while((*p & 0xc0) == 0x80) p++; }
386    
387  /* 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
388  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
389    
# Line 481  Standard C system should have one. */ Line 493  Standard C system should have one. */
493    
494  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
495    
496  /* Private options flags start at the most significant end of the four bytes.  /* Private flags containing information about the compiled regex. They used to
497  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
498  sure they don't overlap! The bits are getting a bit scarce now -- when we run  are in a 16-bit flags word. */
499  out, there is a dummy word in the structure that could be used for the private  
500  bits. */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
501    #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
502  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
503  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
504  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
505  #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 */  
506    
507  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
508    
# Line 507  time, run time, or study time, respectiv Line 518  time, run time, or study time, respectiv
518    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
519     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
520     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| \
521     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
522       PCRE_JAVASCRIPT_COMPAT)
523    
524  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_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_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
527    
528  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
529    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
530     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
531       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
532    
533  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
534    
# Line 596  contain UTF-8 characters with values gre Line 609  contain UTF-8 characters with values gre
609  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
610  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
611  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
612  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
613  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  (which is used for [^] in JavaScript compatibility mode).
614  There are two tests in the code for an escape greater than ESC_b and less than  
615  ESC_Z to detect the types that may be repeated. These are the types that  The final escape must be ESC_REF as subsequent values are used for
616  consume characters. If any new escapes are put in between that don't consume a  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
617  character, that code will have to change. */  greater than ESC_b and less than ESC_Z to detect the types that may be
618    repeated. These are the types that consume characters. If any new escapes are
619    put in between that don't consume a character, that code will have to change.
620    */
621    
622  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,
623         ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
624         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_k, ESC_REF };         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
625           ESC_REF };
626    
627    
628  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
# Line 631  enum { Line 648  enum {
648    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
649    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
650    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
651    OP_ANY,            /* 12 Match any character */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
652    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
653    OP_NOTPROP,        /* 14 \P (not Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
654    OP_PROP,           /* 15 \p (Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
655    OP_ANYNL,          /* 16 \R (any newline sequence) */    OP_PROP,           /* 16 \p (Unicode property) */
656    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */    OP_ANYNL,          /* 17 \R (any newline sequence) */
657    OP_HSPACE,         /* 18 \h (horizontal whitespace) */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
658    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
659    OP_VSPACE,         /* 20 \v (vertical whitespace) */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
660    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
661    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
662    OP_EOD,            /* 23 End of data: \z */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
663      OP_EOD,            /* 24 End of data: \z */
664    OP_OPT,            /* 24 Set runtime options */  
665    OP_CIRC,           /* 25 Start of line - varies with multiline switch */    OP_OPT,            /* 25 Set runtime options */
666    OP_DOLL,           /* 26 End of line - varies with multiline switch */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
667    OP_CHAR,           /* 27 Match one character, casefully */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
668    OP_CHARNC,         /* 28 Match one character, caselessly */    OP_CHAR,           /* 28 Match one character, casefully */
669    OP_NOT,            /* 29 Match one character, not the following one */    OP_CHARNC,         /* 29 Match one character, caselessly */
670      OP_NOT,            /* 30 Match one character, not the following one */
671    OP_STAR,           /* 30 The maximizing and minimizing versions of */  
672    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
673    OP_PLUS,           /* 32 the minimizing one second. */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
674    OP_MINPLUS,        /* 33 This first set applies to single characters.*/    OP_PLUS,           /* 33 the minimizing one second. */
675    OP_QUERY,          /* 34 */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
676    OP_MINQUERY,       /* 35 */    OP_QUERY,          /* 35 */
677      OP_MINQUERY,       /* 36 */
678    OP_UPTO,           /* 36 From 0 to n matches */  
679    OP_MINUPTO,        /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
680    OP_EXACT,          /* 38 Exactly n matches */    OP_MINUPTO,        /* 38 */
681      OP_EXACT,          /* 39 Exactly n matches */
682    OP_POSSTAR,        /* 39 Possessified star */  
683    OP_POSPLUS,        /* 40 Possessified plus */    OP_POSSTAR,        /* 40 Possessified star */
684    OP_POSQUERY,       /* 41 Posesssified query */    OP_POSPLUS,        /* 41 Possessified plus */
685    OP_POSUPTO,        /* 42 Possessified upto */    OP_POSQUERY,       /* 42 Posesssified query */
686      OP_POSUPTO,        /* 43 Possessified upto */
687    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */  
688    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
689    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
690    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
691    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
692    OP_NOTMINQUERY,    /* 48 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
693      OP_NOTMINQUERY,    /* 49 */
694    OP_NOTUPTO,        /* 49 From 0 to n matches */  
695    OP_NOTMINUPTO,     /* 50 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
696    OP_NOTEXACT,       /* 51 Exactly n matches */    OP_NOTMINUPTO,     /* 51 */
697      OP_NOTEXACT,       /* 52 Exactly n matches */
698    OP_NOTPOSSTAR,     /* 52 Possessified versions */  
699    OP_NOTPOSPLUS,     /* 53 */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
700    OP_NOTPOSQUERY,    /* 54 */    OP_NOTPOSPLUS,     /* 54 */
701    OP_NOTPOSUPTO,     /* 55 */    OP_NOTPOSQUERY,    /* 55 */
702      OP_NOTPOSUPTO,     /* 56 */
703    OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */  
704    OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
705    OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
706    OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
707    OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
708    OP_TYPEMINQUERY,   /* 61 */    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
709      OP_TYPEMINQUERY,   /* 62 */
710    OP_TYPEUPTO,       /* 62 From 0 to n matches */  
711    OP_TYPEMINUPTO,    /* 63 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
712    OP_TYPEEXACT,      /* 64 Exactly n matches */    OP_TYPEMINUPTO,    /* 64 */
713      OP_TYPEEXACT,      /* 65 Exactly n matches */
714    OP_TYPEPOSSTAR,    /* 65 Possessified versions */  
715    OP_TYPEPOSPLUS,    /* 66 */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
716    OP_TYPEPOSQUERY,   /* 67 */    OP_TYPEPOSPLUS,    /* 67 */
717    OP_TYPEPOSUPTO,    /* 68 */    OP_TYPEPOSQUERY,   /* 68 */
718      OP_TYPEPOSUPTO,    /* 69 */
719    OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */  
720    OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
721    OP_CRPLUS,         /* 71 the minimizing one second. These codes must */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
722    OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
723    OP_CRQUERY,        /* 73 These are for character classes and back refs */    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
724    OP_CRMINQUERY,     /* 74 */    OP_CRQUERY,        /* 74 These are for character classes and back refs */
725    OP_CRRANGE,        /* 75 These are different to the three sets above. */    OP_CRMINQUERY,     /* 75 */
726    OP_CRMINRANGE,     /* 76 */    OP_CRRANGE,        /* 76 These are different to the three sets above. */
727      OP_CRMINRANGE,     /* 77 */
728    
729    OP_CLASS,          /* 77 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
730    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
731                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
732                             character > 255 is encountered. */                             character > 255 is encountered. */
733    
734    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
735                             class. This does both positive and negative. */                             class. This does both positive and negative. */
736    
737    OP_REF,            /* 80 Match a back reference */    OP_REF,            /* 81 Match a back reference */
738    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
739    OP_CALLOUT,        /* 82 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
740    
741    OP_ALT,            /* 83 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
742    OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */
743    OP_KETRMAX,        /* 85 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
744    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
745    
746    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
747    
748    OP_ASSERT,         /* 87 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
749    OP_ASSERT_NOT,     /* 88 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
750    OP_ASSERTBACK,     /* 89 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
751    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
752    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
753    
754    /* 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,
755    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. */
756    
757    OP_ONCE,           /* 92 Atomic group */    OP_ONCE,           /* 93 Atomic group */
758    OP_BRA,            /* 93 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
759    OP_CBRA,           /* 94 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
760    OP_COND,           /* 95 Conditional group */    OP_COND,           /* 96 Conditional group */
761    
762    /* 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
763    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
764    
765    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
766    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
767    OP_SCOND,          /* 98 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
768    
769    OP_CREF,           /* 99 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
770    OP_RREF,           /* 100 Used to hold a recursion number as condition */    OP_RREF,           /* 101 Used to hold a recursion number as condition */
771    OP_DEF,            /* 101 The DEFINE condition */    OP_DEF,            /* 102 The DEFINE condition */
772    
773    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_BRAZERO,        /* 103 These two must remain together and in this */
774    OP_BRAMINZERO,     /* 103 order. */    OP_BRAMINZERO,     /* 104 order. */
775    
776    /* These are backtracking control verbs */    /* These are backtracking control verbs */
777    
778    OP_PRUNE,          /* 104 */    OP_PRUNE,          /* 105 */
779    OP_SKIP,           /* 105 */    OP_SKIP,           /* 106 */
780    OP_THEN,           /* 106 */    OP_THEN,           /* 107 */
781    OP_COMMIT,         /* 107 */    OP_COMMIT,         /* 108 */
782    
783    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
784    
785    OP_FAIL,           /* 108 */    OP_FAIL,           /* 109 */
786    OP_ACCEPT          /* 109 */    OP_ACCEPT,         /* 110 */
787    
788      /* This is used to skip a subpattern with a {0} quantifier */
789    
790      OP_SKIPZERO        /* 111 */
791  };  };
792    
793    
# Line 774  for debugging. The macro is referenced o Line 796  for debugging. The macro is referenced o
796    
797  #define OP_NAME_LIST \  #define OP_NAME_LIST \
798    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
799    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
800    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
801    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
802    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
# Line 790  for debugging. The macro is referenced o Line 812  for debugging. The macro is referenced o
812    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
813    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
814    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
815    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT"    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
816      "Skip zero"
817    
818    
819  /* 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 806  in UTF-8 mode. The code that uses this t Line 829  in UTF-8 mode. The code that uses this t
829    1,                             /* End                                    */ \    1,                             /* End                                    */ \
830    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
831    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
832    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
833    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
834    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
835    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
# Line 855  in UTF-8 mode. The code that uses this t Line 878  in UTF-8 mode. The code that uses this t
878    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
879    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
880    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
881    1, 1                           /* FAIL, ACCEPT                           */    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
882    
883    
884  /* 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 871  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 894  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
894         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
895         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
896         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
897         ERR60, ERR61 };         ERR60, ERR61, ERR62, ERR63, ERR64 };
898    
899  /* 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
900  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 893  NOTE NOTE NOTE: Line 916  NOTE NOTE NOTE:
916  typedef struct real_pcre {  typedef struct real_pcre {
917    pcre_uint32 magic_number;    pcre_uint32 magic_number;
918    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
919    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
920    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
921      pcre_uint16 dummy1;             /* For future use */
922    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
923    pcre_uint16 top_backref;    pcre_uint16 top_backref;
924    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 934  typedef struct compile_data { Line 957  typedef struct compile_data {
957    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
958    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
959    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
960    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
961      int  final_bracount;          /* Saved value after first pass */
962    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
963    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
964    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
965      int  external_flags;          /* External flag bits to be set */
966    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
   BOOL nopartial;               /* Set TRUE if partial won't work */  
967    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
968    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
969    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
# Line 995  typedef struct match_data { Line 1019  typedef struct match_data {
1019    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1020    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1021    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1022      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1023    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1024    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1025    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1026    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1027      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1028    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1029    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1030    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
# Line 1035  typedef struct dfa_match_data { Line 1061  typedef struct dfa_match_data {
1061  #define ctype_letter  0x02  #define ctype_letter  0x02
1062  #define ctype_digit   0x04  #define ctype_digit   0x04
1063  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1064  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1065  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1066    
1067  /* 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 1063  total length. */ Line 1089  total length. */
1089  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1090    
1091  /* 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
1092  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1093    relocations in shared libraries, it now has an offset into a single string
1094    instead. */
1095    
1096  typedef struct {  typedef struct {
1097    const char *name;    pcre_uint16 name_offset;
1098    pcre_uint16 type;    pcre_uint16 type;
1099    pcre_uint16 value;    pcre_uint16 value;
1100  } ucp_type_table;  } ucp_type_table;
# Line 1084  extern const uschar _pcre_utf8_table4[]; Line 1112  extern const uschar _pcre_utf8_table4[];
1112    
1113  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1114    
1115    extern const char   _pcre_utt_names[];
1116  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1117  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1118    
# Line 1101  extern BOOL         _pcre_is_newline(con Line 1130  extern BOOL         _pcre_is_newline(con
1130  extern int          _pcre_ord2utf8(int, uschar *);  extern int          _pcre_ord2utf8(int, uschar *);
1131  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1132                        const pcre_study_data *, pcre_study_data *);                        const pcre_study_data *, pcre_study_data *);
 extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  
 extern unsigned int _pcre_ucp_othercase(const unsigned int);  
1133  extern int          _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1134  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1135                        int *, BOOL);                        int *, BOOL);
1136  extern BOOL         _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_xclass(int, const uschar *);
1137    
1138    
1139    /* Unicode character database (UCD) */
1140    
1141    typedef struct {
1142      uschar script;
1143      uschar chartype;
1144      pcre_int16 other_case;
1145    } ucd_record;
1146    
1147    extern const ucd_record  ucd_records[];
1148    extern const uschar      ucd_stage1[];
1149    extern const pcre_uint16 ucd_stage2[];
1150    extern const int ucp_gentype[];
1151    
1152    /* UCD access macros */
1153    
1154    #define UCD_BLOCK_SIZE 128
1155    #define GET_UCD(ch) (ucd_records + ucd_stage2[ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1156            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1157    
1158    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1159    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1160    #define UCD_CATEGORY(ch)  ucp_gentype[UCD_CHARTYPE(ch)]
1161    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1162    
1163  #endif  #endif
1164    
1165  /* End of pcre_internal.h */  /* End of pcre_internal.h */

Legend:
Removed from v.213  
changed lines
  Added in v.349

  ViewVC Help
Powered by ViewVC 1.1.5