/[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 178 by ph10, Wed Jun 13 08:44:34 2007 UTC revision 389 by ph10, Sun Mar 15 18:24:05 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 505  time, run time, or study time, respectiv Line 540  time, run time, or study time, respectiv
540  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
541                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
542    
543  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
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       PCRE_NO_START_OPTIMIZE)
554    
555  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
556    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
557     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
558       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
559    
560  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
561    
# Line 542  variable-length repeat, or a anything ot Line 580  variable-length repeat, or a anything ot
580  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
581  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
582    
583  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
584    environments where these macros are defined elsewhere. Unfortunately, there
585    is no way to do the same for the typedef. */
586    
587  typedef int BOOL;  typedef int BOOL;
588    
589    #ifndef FALSE
590  #define FALSE   0  #define FALSE   0
591  #define TRUE    1  #define TRUE    1
592    #endif
593    
594  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
595    
# Line 598  contain UTF-8 characters with values gre Line 640  contain UTF-8 characters with values gre
640  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
641  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
642  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
643  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
644  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  (which is used for [^] in JavaScript compatibility mode).
645  There are two tests in the code for an escape greater than ESC_b and less than  
646  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
647  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
648  character, that code will have to change. */  greater than ESC_b and less than ESC_Z to detect the types that may be
649    repeated. These are the types that consume characters. If any new escapes are
650    put in between that don't consume a character, that code will have to change.
651    */
652    
653  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,
654         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,
655         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,
656           ESC_REF };
657    
658    
659  /* 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  
660  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.
661    
 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.  
   
662  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
663  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
664  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" in pcre_dfa_exec.c that must be updated. */
# Line 638  enum { Line 679  enum {
679    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
680    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
681    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
682    OP_ANY,            /* 12 Match any character */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
683    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
684    OP_NOTPROP,        /* 14 \P (not Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
685    OP_PROP,           /* 15 \p (Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
686    OP_ANYNL,          /* 16 \R (any newline sequence) */    OP_PROP,           /* 16 \p (Unicode property) */
687    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */    OP_ANYNL,          /* 17 \R (any newline sequence) */
688    OP_HSPACE,         /* 18 \h (horizontal whitespace) */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
689    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
690    OP_VSPACE,         /* 20 \v (vertical whitespace) */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
691    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
692    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
693    OP_EOD,            /* 23 End of data: \z */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
694      OP_EOD,            /* 24 End of data: \z */
695    OP_OPT,            /* 24 Set runtime options */  
696    OP_CIRC,           /* 25 Start of line - varies with multiline switch */    OP_OPT,            /* 25 Set runtime options */
697    OP_DOLL,           /* 26 End of line - varies with multiline switch */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
698    OP_CHAR,           /* 27 Match one character, casefully */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
699    OP_CHARNC,         /* 28 Match one character, caselessly */    OP_CHAR,           /* 28 Match one character, casefully */
700    OP_NOT,            /* 29 Match one character, not the following one */    OP_CHARNC,         /* 29 Match one character, caselessly */
701      OP_NOT,            /* 30 Match one character, not the following one */
702    OP_STAR,           /* 30 The maximizing and minimizing versions of */  
703    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
704    OP_PLUS,           /* 32 the minimizing one second. */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
705    OP_MINPLUS,        /* 33 This first set applies to single characters.*/    OP_PLUS,           /* 33 the minimizing one second. */
706    OP_QUERY,          /* 34 */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
707    OP_MINQUERY,       /* 35 */    OP_QUERY,          /* 35 */
708      OP_MINQUERY,       /* 36 */
709    OP_UPTO,           /* 36 From 0 to n matches */  
710    OP_MINUPTO,        /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
711    OP_EXACT,          /* 38 Exactly n matches */    OP_MINUPTO,        /* 38 */
712      OP_EXACT,          /* 39 Exactly n matches */
713    OP_POSSTAR,        /* 39 Possessified star */  
714    OP_POSPLUS,        /* 40 Possessified plus */    OP_POSSTAR,        /* 40 Possessified star */
715    OP_POSQUERY,       /* 41 Posesssified query */    OP_POSPLUS,        /* 41 Possessified plus */
716    OP_POSUPTO,        /* 42 Possessified upto */    OP_POSQUERY,       /* 42 Posesssified query */
717      OP_POSUPTO,        /* 43 Possessified upto */
718    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */  
719    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
720    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
721    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
722    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
723    OP_NOTMINQUERY,    /* 48 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
724      OP_NOTMINQUERY,    /* 49 */
725    OP_NOTUPTO,        /* 49 From 0 to n matches */  
726    OP_NOTMINUPTO,     /* 50 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
727    OP_NOTEXACT,       /* 51 Exactly n matches */    OP_NOTMINUPTO,     /* 51 */
728      OP_NOTEXACT,       /* 52 Exactly n matches */
729    OP_NOTPOSSTAR,     /* 52 Possessified versions */  
730    OP_NOTPOSPLUS,     /* 53 */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
731    OP_NOTPOSQUERY,    /* 54 */    OP_NOTPOSPLUS,     /* 54 */
732    OP_NOTPOSUPTO,     /* 55 */    OP_NOTPOSQUERY,    /* 55 */
733      OP_NOTPOSUPTO,     /* 56 */
734    OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */  
735    OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
736    OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
737    OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
738    OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
739    OP_TYPEMINQUERY,   /* 61 */    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
740      OP_TYPEMINQUERY,   /* 62 */
741    OP_TYPEUPTO,       /* 62 From 0 to n matches */  
742    OP_TYPEMINUPTO,    /* 63 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
743    OP_TYPEEXACT,      /* 64 Exactly n matches */    OP_TYPEMINUPTO,    /* 64 */
744      OP_TYPEEXACT,      /* 65 Exactly n matches */
745    OP_TYPEPOSSTAR,    /* 65 Possessified versions */  
746    OP_TYPEPOSPLUS,    /* 66 */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
747    OP_TYPEPOSQUERY,   /* 67 */    OP_TYPEPOSPLUS,    /* 67 */
748    OP_TYPEPOSUPTO,    /* 68 */    OP_TYPEPOSQUERY,   /* 68 */
749      OP_TYPEPOSUPTO,    /* 69 */
750    OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */  
751    OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
752    OP_CRPLUS,         /* 71 the minimizing one second. These codes must */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
753    OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
754    OP_CRQUERY,        /* 73 These are for character classes and back refs */    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
755    OP_CRMINQUERY,     /* 74 */    OP_CRQUERY,        /* 74 These are for character classes and back refs */
756    OP_CRRANGE,        /* 75 These are different to the three sets above. */    OP_CRMINQUERY,     /* 75 */
757    OP_CRMINRANGE,     /* 76 */    OP_CRRANGE,        /* 76 These are different to the three sets above. */
758      OP_CRMINRANGE,     /* 77 */
759    
760    OP_CLASS,          /* 77 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
761    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
762                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
763                             character > 255 is encountered. */                             character > 255 is encountered. */
764    
765    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
766                             class. This does both positive and negative. */                             class. This does both positive and negative. */
767    
768    OP_REF,            /* 80 Match a back reference */    OP_REF,            /* 81 Match a back reference */
769    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
770    OP_CALLOUT,        /* 82 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
771    
772    OP_ALT,            /* 83 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
773    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 */
774    OP_KETRMAX,        /* 85 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
775    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
776    
777    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
778    
779    OP_ASSERT,         /* 87 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
780    OP_ASSERT_NOT,     /* 88 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
781    OP_ASSERTBACK,     /* 89 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
782    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
783    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
784    
785    /* 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,
786    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. */
787    
788    OP_ONCE,           /* 92 Atomic group */    OP_ONCE,           /* 93 Atomic group */
789    OP_BRA,            /* 83 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
790    OP_CBRA,           /* 94 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
791    OP_COND,           /* 95 Conditional group */    OP_COND,           /* 96 Conditional group */
792    
793    /* 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
794    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
795    
796    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
797    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
798    OP_SCOND,          /* 98 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
799    
800    OP_CREF,           /* 99 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
801    OP_RREF,           /* 100 Used to hold a recursion number as condition */    OP_RREF,           /* 101 Used to hold a recursion number as condition */
802    OP_DEF,            /* 101 The DEFINE condition */    OP_DEF,            /* 102 The DEFINE condition */
803    
804      OP_BRAZERO,        /* 103 These two must remain together and in this */
805      OP_BRAMINZERO,     /* 104 order. */
806    
807      /* These are backtracking control verbs */
808    
809      OP_PRUNE,          /* 105 */
810      OP_SKIP,           /* 106 */
811      OP_THEN,           /* 107 */
812      OP_COMMIT,         /* 108 */
813    
814    OP_BRAZERO,        /* 102 These two must remain together and in this */    /* These are forced failure and success verbs */
815    OP_BRAMINZERO      /* 103 order. */  
816      OP_FAIL,           /* 109 */
817      OP_ACCEPT,         /* 110 */
818    
819      /* This is used to skip a subpattern with a {0} quantifier */
820    
821      OP_SKIPZERO        /* 111 */
822  };  };
823    
824    
# Line 769  for debugging. The macro is referenced o Line 827  for debugging. The macro is referenced o
827    
828  #define OP_NAME_LIST \  #define OP_NAME_LIST \
829    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
830    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
831    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
832    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
833    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
# Line 783  for debugging. The macro is referenced o Line 841  for debugging. The macro is referenced o
841    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
842    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
843    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
844    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
845    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
846      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
847      "Skip zero"
848    
849    
850  /* 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 800  in UTF-8 mode. The code that uses this t Line 860  in UTF-8 mode. The code that uses this t
860    1,                             /* End                                    */ \    1,                             /* End                                    */ \
861    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
862    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
863    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
864    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
865    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
866    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
# Line 848  in UTF-8 mode. The code that uses this t Line 908  in UTF-8 mode. The code that uses this t
908    3,                             /* RREF                                   */ \    3,                             /* RREF                                   */ \
909    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
910    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
911      1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
912      1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
913    
914    
915  /* 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 862  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 924  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
924         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
925         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
926         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
927         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
928           ERR60, ERR61, ERR62, ERR63, ERR64 };
929    
930  /* 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
931  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 884  NOTE NOTE NOTE: Line 947  NOTE NOTE NOTE:
947  typedef struct real_pcre {  typedef struct real_pcre {
948    pcre_uint32 magic_number;    pcre_uint32 magic_number;
949    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
950    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
951    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
952      pcre_uint16 dummy1;             /* For future use */
953    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
954    pcre_uint16 top_backref;    pcre_uint16 top_backref;
955    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 925  typedef struct compile_data { Line 988  typedef struct compile_data {
988    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
989    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
990    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
991    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
992      int  final_bracount;          /* Saved value after first pass */
993    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
994    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
995    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
996      int  external_flags;          /* External flag bits to be set */
997    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
998    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
999    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1000    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1001    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 985  typedef struct match_data { Line 1050  typedef struct match_data {
1050    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1051    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1052    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1053      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1054    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1055    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1056    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1057    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1058      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1059    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1060    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1061    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
# Line 1025  typedef struct dfa_match_data { Line 1092  typedef struct dfa_match_data {
1092  #define ctype_letter  0x02  #define ctype_letter  0x02
1093  #define ctype_digit   0x04  #define ctype_digit   0x04
1094  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1095  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1096  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1097    
1098  /* 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 1053  total length. */ Line 1120  total length. */
1120  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1121    
1122  /* 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
1123  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1124    relocations in shared libraries, it now has an offset into a single string
1125    instead. */
1126    
1127  typedef struct {  typedef struct {
1128    const char *name;    pcre_uint16 name_offset;
1129    pcre_uint16 type;    pcre_uint16 type;
1130    pcre_uint16 value;    pcre_uint16 value;
1131  } ucp_type_table;  } ucp_type_table;
# Line 1074  extern const uschar _pcre_utf8_table4[]; Line 1143  extern const uschar _pcre_utf8_table4[];
1143    
1144  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1145    
1146    extern const char   _pcre_utt_names[];
1147  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1148  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1149    
# Line 1091  extern BOOL _pcre_is_newline(con Line 1161  extern BOOL _pcre_is_newline(con
1161  extern int          _pcre_ord2utf8(int, uschar *);  extern int          _pcre_ord2utf8(int, uschar *);
1162  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1163                        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);  
1164  extern int          _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1165  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1166                        int *, BOOL);                        int *, BOOL);
1167  extern BOOL         _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_xclass(int, const uschar *);
1168    
1169    
1170    /* Unicode character database (UCD) */
1171    
1172    typedef struct {
1173      uschar script;
1174      uschar chartype;
1175      pcre_int32 other_case;
1176    } ucd_record;
1177    
1178    extern const ucd_record  _pcre_ucd_records[];
1179    extern const uschar      _pcre_ucd_stage1[];
1180    extern const pcre_uint16 _pcre_ucd_stage2[];
1181    extern const int         _pcre_ucp_gentype[];
1182    
1183    
1184    /* UCD access macros */
1185    
1186    #define UCD_BLOCK_SIZE 128
1187    #define GET_UCD(ch) (_pcre_ucd_records + \
1188            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1189            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1190    
1191    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1192    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1193    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1194    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1195    
1196  #endif  #endif
1197    
1198  /* End of pcre_internal.h */  /* End of pcre_internal.h */

Legend:
Removed from v.178  
changed lines
  Added in v.389

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12