/[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 166 by ph10, Wed May 9 14:48:28 2007 UTC revision 384 by ph10, Sun Mar 8 16:27:43 2009 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) 1997-2009 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 137  PCRE_EXP_DATA_DEFN only if they are not Line 132  PCRE_EXP_DATA_DEFN only if they are not
132  #  endif  #  endif
133  #endif  #endif
134    
135    /* When compiling with the MSVC compiler, it is sometimes necessary to include
136    a "calling convention" before exported function names. (This is secondhand
137    information; I know nothing about MSVC myself). For example, something like
138    
139      void __cdecl function(....)
140    
141    might be needed. In order so make this easy, all the exported functions have
142    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
143    set, we ensure here that it has no effect. */
144    
145    #ifndef PCRE_CALL_CONVENTION
146    #define PCRE_CALL_CONVENTION
147    #endif
148    
149  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
150  cannot determine these outside the compilation (e.g. by running a program as  cannot determine these outside the compilation (e.g. by running a program as
151  part of "configure") because PCRE is often cross-compiled for use on other  part of "configure") because PCRE is often cross-compiled for use on other
# Line 145  preprocessor time in standard C environm Line 154  preprocessor time in standard C environm
154    
155  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
156    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
157      typedef short pcre_int16;
158  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
159    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
160      typedef int pcre_int16;
161  #else  #else
162    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
163  #endif  #endif
164    
165  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
166    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
167      typedef int pcre_int32;
168  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
169    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
170      typedef long int pcre_int32;
171  #else  #else
172    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
173  #endif  #endif
# Line 234  must begin with PCRE_. */ Line 247  must begin with PCRE_. */
247  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
248  values. */  values. */
249    
250  #include <pcre.h>  #include "pcre.h"
251  #include "ucp.h"  #include "ucp.h"
252    
253  /* 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 376  capturing parenthesis numbers in back re
376    
377  /* 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
378  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
379  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
380    never be called in byte mode. To make sure it can never even appear when UTF-8
381    support is omitted, we don't even define it. */
382    
383  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
384  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 371  byte-mode, and more complicated ones for Line 386  byte-mode, and more complicated ones for
386  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
387  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
388  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
389  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
390    
391  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
392    
# Line 463  if there are extra bytes. This is called Line 478  if there are extra bytes. This is called
478      len += gcaa; \      len += gcaa; \
479      }      }
480    
481    /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
482    pointer, incrementing length if there are extra bytes. This is called when we
483    know we are in UTF-8 mode. */
484    
485    #define GETCHARLENTEST(c, eptr, len) \
486      c = *eptr; \
487      if (utf8 && c >= 0xc0) \
488        { \
489        int gcii; \
490        int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
491        int gcss = 6*gcaa; \
492        c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
493        for (gcii = 1; gcii <= gcaa; gcii++) \
494          { \
495          gcss -= 6; \
496          c |= (eptr[gcii] & 0x3f) << gcss; \
497          } \
498        len += gcaa; \
499        }
500    
501  /* 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
502  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
503    because almost all calls are already within a block of UTF-8 only code. */
504    
505  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
506    
507  #endif  #endif
508    
# Line 483  Standard C system should have one. */ Line 519  Standard C system should have one. */
519    
520  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
521    
522  /* Private options flags start at the most significant end of the four bytes.  /* Private flags containing information about the compiled regex. They used to
523  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
524  sure they don't overlap! The bits are getting a bit scarce now -- when we run  are in a 16-bit flags word. */
525  out, there is a dummy word in the structure that could be used for the private  
526  bits. */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
527    #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
528  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
529  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
530  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
531  #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 */  
532    
533  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
534    
# Line 509  time, run time, or study time, respectiv Line 544  time, run time, or study time, respectiv
544    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
545     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
546     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| \
547     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
548       PCRE_JAVASCRIPT_COMPAT)
549    
550  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
551    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
552     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
553    
554  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
555    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
556     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
557       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
558    
559  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
560    
# Line 542  variable-length repeat, or a anything ot Line 579  variable-length repeat, or a anything ot
579  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
580  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
581    
582  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
583    environments where these macros are defined elsewhere. Unfortunately, there
584    is no way to do the same for the typedef. */
585    
586  typedef int BOOL;  typedef int BOOL;
587    
588    #ifndef FALSE
589  #define FALSE   0  #define FALSE   0
590  #define TRUE    1  #define TRUE    1
591    #endif
592    
593  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
594    
# Line 598  contain UTF-8 characters with values gre Line 639  contain UTF-8 characters with values gre
639  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
640  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
641  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
642  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
643  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_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 };  
644    
645    The final escape must be ESC_REF as subsequent values are used for
646    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
647    greater than ESC_b and less than ESC_Z to detect the types that may be
648    repeated. These are the types that consume characters. If any new escapes are
649    put in between that don't consume a character, that code will have to change.
650    */
651    
652  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
653  that extract substrings. Starting from 1 (i.e. after OP_END), the values up to         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
654  OP_EOD must correspond in order to the list of escapes immediately above.         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
655           ESC_REF };
656    
657  To keep stored, compiled patterns compatible, new opcodes should be added  
658  immediately before OP_BRA, where (since release 7.0) a gap is left for this  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
659  purpose.  OP_EOD must correspond in order to the list of escapes immediately above.
660    
661  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
662  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 669  enum {
669    
670    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
671    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
672    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
673    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
674    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
675    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
676    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
677    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
678    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
679    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
680    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
681    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
682    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
683    OP_PROP,           /* 14 \p (Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
684    OP_ANYNL,          /* 15 \R (any newline sequence) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
685    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */    OP_PROP,           /* 16 \p (Unicode property) */
686    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */    OP_ANYNL,          /* 17 \R (any newline sequence) */
687    OP_EOD,            /* 18 End of data: \z */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
688      OP_HSPACE,         /* 19 \h (horizontal whitespace) */
689    OP_OPT,            /* 19 Set runtime options */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
690    OP_CIRC,           /* 20 Start of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
691    OP_DOLL,           /* 21 End of line - varies with multiline switch */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
692    OP_CHAR,           /* 22 Match one character, casefully */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
693    OP_CHARNC,         /* 23 Match one character, caselessly */    OP_EOD,            /* 24 End of data: \z */
694    OP_NOT,            /* 24 Match one character, not the following one */  
695      OP_OPT,            /* 25 Set runtime options */
696    OP_STAR,           /* 25 The maximizing and minimizing versions of */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
697    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
698    OP_PLUS,           /* 27 the minimizing one second. */    OP_CHAR,           /* 28 Match one character, casefully */
699    OP_MINPLUS,        /* 28 This first set applies to single characters.*/    OP_CHARNC,         /* 29 Match one character, caselessly */
700    OP_QUERY,          /* 29 */    OP_NOT,            /* 30 Match one character, not the following one */
701    OP_MINQUERY,       /* 30 */  
702      OP_STAR,           /* 31 The maximizing and minimizing versions of */
703    OP_UPTO,           /* 31 From 0 to n matches */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
704    OP_MINUPTO,        /* 32 */    OP_PLUS,           /* 33 the minimizing one second. */
705    OP_EXACT,          /* 33 Exactly n matches */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
706      OP_QUERY,          /* 35 */
707    OP_POSSTAR,        /* 34 Possessified star */    OP_MINQUERY,       /* 36 */
708    OP_POSPLUS,        /* 35 Possessified plus */  
709    OP_POSQUERY,       /* 36 Posesssified query */    OP_UPTO,           /* 37 From 0 to n matches */
710    OP_POSUPTO,        /* 37 Possessified upto */    OP_MINUPTO,        /* 38 */
711      OP_EXACT,          /* 39 Exactly n matches */
712    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */  
713    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */    OP_POSSTAR,        /* 40 Possessified star */
714    OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */    OP_POSPLUS,        /* 41 Possessified plus */
715    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */    OP_POSQUERY,       /* 42 Posesssified query */
716    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */    OP_POSUPTO,        /* 43 Possessified upto */
717    OP_NOTMINQUERY,    /* 43 */  
718      OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
719    OP_NOTUPTO,        /* 44 From 0 to n matches */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
720    OP_NOTMINUPTO,     /* 45 */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
721    OP_NOTEXACT,       /* 46 Exactly n matches */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
722      OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
723    OP_NOTPOSSTAR,     /* 47 Possessified versions */    OP_NOTMINQUERY,    /* 49 */
724    OP_NOTPOSPLUS,     /* 48 */  
725    OP_NOTPOSQUERY,    /* 49 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
726    OP_NOTPOSUPTO,     /* 50 */    OP_NOTMINUPTO,     /* 51 */
727      OP_NOTEXACT,       /* 52 Exactly n matches */
728    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */  
729    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
730    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */    OP_NOTPOSPLUS,     /* 54 */
731    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */    OP_NOTPOSQUERY,    /* 55 */
732    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */    OP_NOTPOSUPTO,     /* 56 */
733    OP_TYPEMINQUERY,   /* 56 */  
734      OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
735    OP_TYPEUPTO,       /* 57 From 0 to n matches */    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
736    OP_TYPEMINUPTO,    /* 58 */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
737    OP_TYPEEXACT,      /* 59 Exactly n matches */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
738      OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
739    OP_TYPEPOSSTAR,    /* 60 Possessified versions */    OP_TYPEMINQUERY,   /* 62 */
740    OP_TYPEPOSPLUS,    /* 61 */  
741    OP_TYPEPOSQUERY,   /* 62 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
742    OP_TYPEPOSUPTO,    /* 63 */    OP_TYPEMINUPTO,    /* 64 */
743      OP_TYPEEXACT,      /* 65 Exactly n matches */
744    OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */  
745    OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
746    OP_CRPLUS,         /* 66 the minimizing one second. These codes must */    OP_TYPEPOSPLUS,    /* 67 */
747    OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */    OP_TYPEPOSQUERY,   /* 68 */
748    OP_CRQUERY,        /* 68 These are for character classes and back refs */    OP_TYPEPOSUPTO,    /* 69 */
749    OP_CRMINQUERY,     /* 69 */  
750    OP_CRRANGE,        /* 70 These are different to the three sets above. */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
751    OP_CRMINRANGE,     /* 71 */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
752      OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
753      OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
754      OP_CRQUERY,        /* 74 These are for character classes and back refs */
755      OP_CRMINQUERY,     /* 75 */
756      OP_CRRANGE,        /* 76 These are different to the three sets above. */
757      OP_CRMINRANGE,     /* 77 */
758    
759    OP_CLASS,          /* 72 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
760    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
761                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
762                             character > 255 is encountered. */                             character > 255 is encountered. */
763    
764    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
765                             class. This does both positive and negative. */                             class. This does both positive and negative. */
766    
767    OP_REF,            /* 75 Match a back reference */    OP_REF,            /* 81 Match a back reference */
768    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
769    OP_CALLOUT,        /* 77 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
770    
771    OP_ALT,            /* 78 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
772    OP_KET,            /* 79 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */
773    OP_KETRMAX,        /* 80 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
774    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
775    
776    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
777    
778    OP_ASSERT,         /* 82 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
779    OP_ASSERT_NOT,     /* 83 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
780    OP_ASSERTBACK,     /* 84 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
781    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
782    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
783    
784    /* 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,
785    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. */
786    
787    OP_ONCE,           /* 87 Atomic group */    OP_ONCE,           /* 93 Atomic group */
788    OP_BRA,            /* 88 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
789    OP_CBRA,           /* 89 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
790    OP_COND,           /* 90 Conditional group */    OP_COND,           /* 96 Conditional group */
791    
792    /* 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
793    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
794    
795    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
796    OP_SCBRA,          /* 92 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
797    OP_SCOND,          /* 93 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
798    
799    OP_CREF,           /* 94 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
800    OP_RREF,           /* 95 Used to hold a recursion number as condition */    OP_RREF,           /* 101 Used to hold a recursion number as condition */
801    OP_DEF,            /* 96 The DEFINE condition */    OP_DEF,            /* 102 The DEFINE condition */
802    
803      OP_BRAZERO,        /* 103 These two must remain together and in this */
804      OP_BRAMINZERO,     /* 104 order. */
805    
806      /* These are backtracking control verbs */
807    
808    OP_BRAZERO,        /* 97 These two must remain together and in this */    OP_PRUNE,          /* 105 */
809    OP_BRAMINZERO      /* 98 order. */    OP_SKIP,           /* 106 */
810      OP_THEN,           /* 107 */
811      OP_COMMIT,         /* 108 */
812    
813      /* These are forced failure and success verbs */
814    
815      OP_FAIL,           /* 109 */
816      OP_ACCEPT,         /* 110 */
817    
818      /* This is used to skip a subpattern with a {0} quantifier */
819    
820      OP_SKIPZERO        /* 111 */
821  };  };
822    
823    
# Line 763  enum { Line 825  enum {
825  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
826    
827  #define OP_NAME_LIST \  #define OP_NAME_LIST \
828    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
829    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
830    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
831    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
832    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
833    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
834    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 778  for debugging. The macro is referenced o Line 840  for debugging. The macro is referenced o
840    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
841    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
842    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
843    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
844    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
845      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
846      "Skip zero"
847    
848    
849  /* 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 793  in UTF-8 mode. The code that uses this t Line 857  in UTF-8 mode. The code that uses this t
857    
858  #define OP_LENGTHS \  #define OP_LENGTHS \
859    1,                             /* End                                    */ \    1,                             /* End                                    */ \
860    1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
861    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
862    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
863      3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
864      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
865    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
866    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
867    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 841  in UTF-8 mode. The code that uses this t Line 907  in UTF-8 mode. The code that uses this t
907    3,                             /* RREF                                   */ \    3,                             /* RREF                                   */ \
908    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
909    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
910      1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
911      1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
912    
913    
914  /* 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 855  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 923  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
923         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
924         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
925         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
926         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
927           ERR60, ERR61, ERR62, ERR63, ERR64 };
928    
929  /* 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
930  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 877  NOTE NOTE NOTE: Line 946  NOTE NOTE NOTE:
946  typedef struct real_pcre {  typedef struct real_pcre {
947    pcre_uint32 magic_number;    pcre_uint32 magic_number;
948    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
949    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
950    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
951      pcre_uint16 dummy1;             /* For future use */
952    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
953    pcre_uint16 top_backref;    pcre_uint16 top_backref;
954    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 918  typedef struct compile_data { Line 987  typedef struct compile_data {
987    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
988    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
989    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
990    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
991      int  final_bracount;          /* Saved value after first pass */
992    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
993    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
994    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
995      int  external_flags;          /* External flag bits to be set */
996    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
997    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
998    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
999    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1000    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 944  typedef struct recursion_info { Line 1015  typedef struct recursion_info {
1015    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1016    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1017    const uschar *after_call;     /* "Return value": points after the call in the expr */    const uschar *after_call;     /* "Return value": points after the call in the expr */
1018    USPTR save_start;             /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
1019    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1020    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1021  } recursion_info;  } recursion_info;
# Line 978  typedef struct match_data { Line 1049  typedef struct match_data {
1049    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1050    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1051    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1052      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1053    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1054    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1055    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1056    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1057      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1058    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1059    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1060    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1061    USPTR  start_match;           /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1062    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1063    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1064    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
# Line 1018  typedef struct dfa_match_data { Line 1091  typedef struct dfa_match_data {
1091  #define ctype_letter  0x02  #define ctype_letter  0x02
1092  #define ctype_digit   0x04  #define ctype_digit   0x04
1093  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1094  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1095  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1096    
1097  /* 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 1046  total length. */ Line 1119  total length. */
1119  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1120    
1121  /* 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
1122  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1123    relocations in shared libraries, it now has an offset into a single string
1124    instead. */
1125    
1126  typedef struct {  typedef struct {
1127    const char *name;    pcre_uint16 name_offset;
1128    pcre_uint16 type;    pcre_uint16 type;
1129    pcre_uint16 value;    pcre_uint16 value;
1130  } ucp_type_table;  } ucp_type_table;
# Line 1067  extern const uschar _pcre_utf8_table4[]; Line 1142  extern const uschar _pcre_utf8_table4[];
1142    
1143  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1144    
1145    extern const char   _pcre_utt_names[];
1146  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1147  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1148    
# Line 1084  extern BOOL _pcre_is_newline(con Line 1160  extern BOOL _pcre_is_newline(con
1160  extern int          _pcre_ord2utf8(int, uschar *);  extern int          _pcre_ord2utf8(int, uschar *);
1161  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1162                        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);  
1163  extern int          _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1164  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1165                        int *, BOOL);                        int *, BOOL);
1166  extern BOOL         _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_xclass(int, const uschar *);
1167    
1168    
1169    /* Unicode character database (UCD) */
1170    
1171    typedef struct {
1172      uschar script;
1173      uschar chartype;
1174      pcre_int32 other_case;
1175    } ucd_record;
1176    
1177    extern const ucd_record  _pcre_ucd_records[];
1178    extern const uschar      _pcre_ucd_stage1[];
1179    extern const pcre_uint16 _pcre_ucd_stage2[];
1180    extern const int         _pcre_ucp_gentype[];
1181    
1182    
1183    /* UCD access macros */
1184    
1185    #define UCD_BLOCK_SIZE 128
1186    #define GET_UCD(ch) (_pcre_ucd_records + \
1187            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1188            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1189    
1190    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1191    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1192    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1193    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1194    
1195  #endif  #endif
1196    
1197  /* End of pcre_internal.h */  /* End of pcre_internal.h */

Legend:
Removed from v.166  
changed lines
  Added in v.384

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12