/[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 172 by ph10, Tue Jun 5 10:40:13 2007 UTC revision 369 by ph10, Sun Aug 24 16:53:47 2008 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2007 University of Cambridge             Copyright (c) 1997-2008 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 67  be absolutely sure we get our version. * Line 67  be absolutely sure we get our version. *
67  #endif  #endif
68    
69    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
70  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
71  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
72    
# Line 112  PCRE_EXP_DATA_DEFN only if they are not Line 108  PCRE_EXP_DATA_DEFN only if they are not
108    
109  #ifndef PCRE_EXP_DECL  #ifndef PCRE_EXP_DECL
110  #  ifdef _WIN32  #  ifdef _WIN32
111  #    ifdef DLL_EXPORT  #    ifndef PCRE_STATIC
112  #      define PCRE_EXP_DECL       extern __declspec(dllexport)  #      define PCRE_EXP_DECL       extern __declspec(dllexport)
113  #      define PCRE_EXP_DEFN       __declspec(dllexport)  #      define PCRE_EXP_DEFN       __declspec(dllexport)
114  #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)  #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
# Line 121  PCRE_EXP_DATA_DEFN only if they are not Line 117  PCRE_EXP_DATA_DEFN only if they are not
117  #      define PCRE_EXP_DEFN  #      define PCRE_EXP_DEFN
118  #      define PCRE_EXP_DATA_DEFN  #      define PCRE_EXP_DATA_DEFN
119  #    endif  #    endif
 #  
120  #  else  #  else
121  #    ifdef __cplusplus  #    ifdef __cplusplus
122  #      define PCRE_EXP_DECL       extern "C"  #      define PCRE_EXP_DECL       extern "C"
# Line 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 464  if there are extra bytes. This is called Line 479  if there are extra bytes. This is called
479      }      }
480    
481  /* 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
482  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
483    because almost all calls are already within a block of UTF-8 only code. */
484    
485  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
486    
487  #endif  #endif
488    
# Line 483  Standard C system should have one. */ Line 499  Standard C system should have one. */
499    
500  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
501    
502  /* Private options flags start at the most significant end of the four bytes.  /* Private flags containing information about the compiled regex. They used to
503  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
504  sure they don't overlap! The bits are getting a bit scarce now -- when we run  are in a 16-bit flags word. */
505  out, there is a dummy word in the structure that could be used for the private  
506  bits. */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
507    #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
508  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
509  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
510  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
511  #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 */  
512    
513  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
514    
# Line 509  time, run time, or study time, respectiv Line 524  time, run time, or study time, respectiv
524    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
525     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
526     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| \
527     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
528       PCRE_JAVASCRIPT_COMPAT)
529    
530  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
531    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
532     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
533    
534  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
535    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
536     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
537       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
538    
539  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
540    
# Line 542  variable-length repeat, or a anything ot Line 559  variable-length repeat, or a anything ot
559  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
560  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
561    
562  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
563    environments where these macros are defined elsewhere. */
564    
565    #ifndef FALSE
566  typedef int BOOL;  typedef int BOOL;
567    
568  #define FALSE   0  #define FALSE   0
569  #define TRUE    1  #define TRUE    1
570    #endif
571    
572  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
573    
# Line 598  contain UTF-8 characters with values gre Line 618  contain UTF-8 characters with values gre
618  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
619  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
620  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
621  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
622  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  (which is used for [^] in JavaScript compatibility mode).
623  There are two tests in the code for an escape greater than ESC_b and less than  
624  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
625  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
626  character, that code will have to change. */  greater than ESC_b and less than ESC_Z to detect the types that may be
627    repeated. These are the types that consume characters. If any new escapes are
628    put in between that don't consume a character, that code will have to change.
629    */
630    
631  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,
632         ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
633         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,
634           ESC_REF };
635    
636    
637  /* 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  
638  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.
639    
 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.  
   
640  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
641  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
642  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" in pcre_dfa_exec.c that must be updated. */
# Line 638  enum { Line 657  enum {
657    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
658    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
659    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
660    OP_ANY,            /* 12 Match any character */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
661    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
662    OP_NOTPROP,        /* 14 \P (not Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
663    OP_PROP,           /* 15 \p (Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
664    OP_ANYNL,          /* 16 \R (any newline sequence) */    OP_PROP,           /* 16 \p (Unicode property) */
665    OP_EXTUNI,         /* 17 \X (extended Unicode sequence */    OP_ANYNL,          /* 17 \R (any newline sequence) */
666    OP_EODN,           /* 18 End of data or \n at end of data: \Z. */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
667    OP_EOD,            /* 19 End of data: \z */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
668      OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
669    OP_OPT,            /* 20 Set runtime options */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
670    OP_CIRC,           /* 21 Start of line - varies with multiline switch */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
671    OP_DOLL,           /* 22 End of line - varies with multiline switch */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
672    OP_CHAR,           /* 23 Match one character, casefully */    OP_EOD,            /* 24 End of data: \z */
673    OP_CHARNC,         /* 24 Match one character, caselessly */  
674    OP_NOT,            /* 25 Match one character, not the following one */    OP_OPT,            /* 25 Set runtime options */
675      OP_CIRC,           /* 26 Start of line - varies with multiline switch */
676    OP_STAR,           /* 26 The maximizing and minimizing versions of */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
677    OP_MINSTAR,        /* 27 these six opcodes must come in pairs, with */    OP_CHAR,           /* 28 Match one character, casefully */
678    OP_PLUS,           /* 28 the minimizing one second. */    OP_CHARNC,         /* 29 Match one character, caselessly */
679    OP_MINPLUS,        /* 29 This first set applies to single characters.*/    OP_NOT,            /* 30 Match one character, not the following one */
680    OP_QUERY,          /* 30 */  
681    OP_MINQUERY,       /* 31 */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
682      OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
683    OP_UPTO,           /* 32 From 0 to n matches */    OP_PLUS,           /* 33 the minimizing one second. */
684    OP_MINUPTO,        /* 33 */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
685    OP_EXACT,          /* 34 Exactly n matches */    OP_QUERY,          /* 35 */
686      OP_MINQUERY,       /* 36 */
687    OP_POSSTAR,        /* 35 Possessified star */  
688    OP_POSPLUS,        /* 36 Possessified plus */    OP_UPTO,           /* 37 From 0 to n matches */
689    OP_POSQUERY,       /* 37 Posesssified query */    OP_MINUPTO,        /* 38 */
690    OP_POSUPTO,        /* 38 Possessified upto */    OP_EXACT,          /* 39 Exactly n matches */
691    
692    OP_NOTSTAR,        /* 39 The maximizing and minimizing versions of */    OP_POSSTAR,        /* 40 Possessified star */
693    OP_NOTMINSTAR,     /* 40 these six opcodes must come in pairs, with */    OP_POSPLUS,        /* 41 Possessified plus */
694    OP_NOTPLUS,        /* 41 the minimizing one second. They must be in */    OP_POSQUERY,       /* 42 Posesssified query */
695    OP_NOTMINPLUS,     /* 42 exactly the same order as those above. */    OP_POSUPTO,        /* 43 Possessified upto */
696    OP_NOTQUERY,       /* 43 This set applies to "not" single characters. */  
697    OP_NOTMINQUERY,    /* 44 */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
698      OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
699    OP_NOTUPTO,        /* 45 From 0 to n matches */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
700    OP_NOTMINUPTO,     /* 46 */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
701    OP_NOTEXACT,       /* 47 Exactly n matches */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
702      OP_NOTMINQUERY,    /* 49 */
703    OP_NOTPOSSTAR,     /* 48 Possessified versions */  
704    OP_NOTPOSPLUS,     /* 49 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
705    OP_NOTPOSQUERY,    /* 50 */    OP_NOTMINUPTO,     /* 51 */
706    OP_NOTPOSUPTO,     /* 51 */    OP_NOTEXACT,       /* 52 Exactly n matches */
707    
708    OP_TYPESTAR,       /* 52 The maximizing and minimizing versions of */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
709    OP_TYPEMINSTAR,    /* 53 these six opcodes must come in pairs, with */    OP_NOTPOSPLUS,     /* 54 */
710    OP_TYPEPLUS,       /* 54 the minimizing one second. These codes must */    OP_NOTPOSQUERY,    /* 55 */
711    OP_TYPEMINPLUS,    /* 55 be in exactly the same order as those above. */    OP_NOTPOSUPTO,     /* 56 */
712    OP_TYPEQUERY,      /* 56 This set applies to character types such as \d */  
713    OP_TYPEMINQUERY,   /* 57 */    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
714      OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
715    OP_TYPEUPTO,       /* 58 From 0 to n matches */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
716    OP_TYPEMINUPTO,    /* 59 */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
717    OP_TYPEEXACT,      /* 60 Exactly n matches */    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
718      OP_TYPEMINQUERY,   /* 62 */
719    OP_TYPEPOSSTAR,    /* 61 Possessified versions */  
720    OP_TYPEPOSPLUS,    /* 62 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
721    OP_TYPEPOSQUERY,   /* 63 */    OP_TYPEMINUPTO,    /* 64 */
722    OP_TYPEPOSUPTO,    /* 64 */    OP_TYPEEXACT,      /* 65 Exactly n matches */
723    
724    OP_CRSTAR,         /* 65 The maximizing and minimizing versions of */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
725    OP_CRMINSTAR,      /* 66 all these opcodes must come in pairs, with */    OP_TYPEPOSPLUS,    /* 67 */
726    OP_CRPLUS,         /* 67 the minimizing one second. These codes must */    OP_TYPEPOSQUERY,   /* 68 */
727    OP_CRMINPLUS,      /* 68 be in exactly the same order as those above. */    OP_TYPEPOSUPTO,    /* 69 */
728    OP_CRQUERY,        /* 69 These are for character classes and back refs */  
729    OP_CRMINQUERY,     /* 70 */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
730    OP_CRRANGE,        /* 71 These are different to the three sets above. */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
731    OP_CRMINRANGE,     /* 72 */    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
732      OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
733      OP_CRQUERY,        /* 74 These are for character classes and back refs */
734      OP_CRMINQUERY,     /* 75 */
735      OP_CRRANGE,        /* 76 These are different to the three sets above. */
736      OP_CRMINRANGE,     /* 77 */
737    
738    OP_CLASS,          /* 73 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
739    OP_NCLASS,         /* 74 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
740                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
741                             character > 255 is encountered. */                             character > 255 is encountered. */
742    
743    OP_XCLASS,         /* 75 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
744                             class. This does both positive and negative. */                             class. This does both positive and negative. */
745    
746    OP_REF,            /* 76 Match a back reference */    OP_REF,            /* 81 Match a back reference */
747    OP_RECURSE,        /* 77 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
748    OP_CALLOUT,        /* 78 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
749    
750    OP_ALT,            /* 79 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
751    OP_KET,            /* 80 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */
752    OP_KETRMAX,        /* 81 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
753    OP_KETRMIN,        /* 82 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
754    
755    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
756    
757    OP_ASSERT,         /* 83 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
758    OP_ASSERT_NOT,     /* 84 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
759    OP_ASSERTBACK,     /* 85 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
760    OP_ASSERTBACK_NOT, /* 86 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
761    OP_REVERSE,        /* 87 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
762    
763    /* 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,
764    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. */
765    
766    OP_ONCE,           /* 88 Atomic group */    OP_ONCE,           /* 93 Atomic group */
767    OP_BRA,            /* 89 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
768    OP_CBRA,           /* 90 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
769    OP_COND,           /* 91 Conditional group */    OP_COND,           /* 96 Conditional group */
770    
771    /* 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
772    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
773    
774    OP_SBRA,           /* 92 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
775    OP_SCBRA,          /* 93 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
776    OP_SCOND,          /* 94 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
777    
778    OP_CREF,           /* 95 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
779    OP_RREF,           /* 96 Used to hold a recursion number as condition */    OP_RREF,           /* 101 Used to hold a recursion number as condition */
780    OP_DEF,            /* 97 The DEFINE condition */    OP_DEF,            /* 102 The DEFINE condition */
781    
782      OP_BRAZERO,        /* 103 These two must remain together and in this */
783      OP_BRAMINZERO,     /* 104 order. */
784    
785    OP_BRAZERO,        /* 98 These two must remain together and in this */    /* These are backtracking control verbs */
786    OP_BRAMINZERO      /* 99 order. */  
787      OP_PRUNE,          /* 105 */
788      OP_SKIP,           /* 106 */
789      OP_THEN,           /* 107 */
790      OP_COMMIT,         /* 108 */
791    
792      /* These are forced failure and success verbs */
793    
794      OP_FAIL,           /* 109 */
795      OP_ACCEPT,         /* 110 */
796    
797      /* This is used to skip a subpattern with a {0} quantifier */
798    
799      OP_SKIPZERO        /* 111 */
800  };  };
801    
802    
# Line 765  for debugging. The macro is referenced o Line 805  for debugging. The macro is referenced o
805    
806  #define OP_NAME_LIST \  #define OP_NAME_LIST \
807    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
808    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
809    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
810    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
811    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
812    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
813    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 779  for debugging. The macro is referenced o Line 819  for debugging. The macro is referenced o
819    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
820    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
821    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
822    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
823    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
824      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
825      "Skip zero"
826    
827    
828  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 796  in UTF-8 mode. The code that uses this t Line 838  in UTF-8 mode. The code that uses this t
838    1,                             /* End                                    */ \    1,                             /* End                                    */ \
839    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
840    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
841    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
842    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
843      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
844    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
845    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
846    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 843  in UTF-8 mode. The code that uses this t Line 886  in UTF-8 mode. The code that uses this t
886    3,                             /* RREF                                   */ \    3,                             /* RREF                                   */ \
887    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
888    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
889      1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
890      1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
891    
892    
893  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
# Line 857  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 902  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
902         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
903         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
904         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
905         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
906           ERR60, ERR61, ERR62, ERR63, ERR64 };
907    
908  /* 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
909  code vector run on as long as necessary after the end. We store an explicit  code vector run on as long as necessary after the end. We store an explicit
# Line 879  NOTE NOTE NOTE: Line 925  NOTE NOTE NOTE:
925  typedef struct real_pcre {  typedef struct real_pcre {
926    pcre_uint32 magic_number;    pcre_uint32 magic_number;
927    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
928    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
929    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
930      pcre_uint16 dummy1;             /* For future use */
931    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
932    pcre_uint16 top_backref;    pcre_uint16 top_backref;
933    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 920  typedef struct compile_data { Line 966  typedef struct compile_data {
966    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
967    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
968    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
969    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
970      int  final_bracount;          /* Saved value after first pass */
971    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
972    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
973    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
974      int  external_flags;          /* External flag bits to be set */
975    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
976    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
977    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
978    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
979    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 980  typedef struct match_data { Line 1028  typedef struct match_data {
1028    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1029    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1030    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1031      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1032    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1033    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1034    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1035    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1036      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1037    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1038    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1039    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
# Line 1020  typedef struct dfa_match_data { Line 1070  typedef struct dfa_match_data {
1070  #define ctype_letter  0x02  #define ctype_letter  0x02
1071  #define ctype_digit   0x04  #define ctype_digit   0x04
1072  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1073  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1074  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1075    
1076  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
# Line 1048  total length. */ Line 1098  total length. */
1098  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1099    
1100  /* 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
1101  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1102    relocations in shared libraries, it now has an offset into a single string
1103    instead. */
1104    
1105  typedef struct {  typedef struct {
1106    const char *name;    pcre_uint16 name_offset;
1107    pcre_uint16 type;    pcre_uint16 type;
1108    pcre_uint16 value;    pcre_uint16 value;
1109  } ucp_type_table;  } ucp_type_table;
# Line 1069  extern const uschar _pcre_utf8_table4[]; Line 1121  extern const uschar _pcre_utf8_table4[];
1121    
1122  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1123    
1124    extern const char   _pcre_utt_names[];
1125  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1126  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1127    
# Line 1086  extern BOOL _pcre_is_newline(con Line 1139  extern BOOL _pcre_is_newline(con
1139  extern int          _pcre_ord2utf8(int, uschar *);  extern int          _pcre_ord2utf8(int, uschar *);
1140  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1141                        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);  
1142  extern int          _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1143  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1144                        int *, BOOL);                        int *, BOOL);
1145  extern BOOL         _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_xclass(int, const uschar *);
1146    
1147    
1148    /* Unicode character database (UCD) */
1149    
1150    typedef struct {
1151      uschar script;
1152      uschar chartype;
1153      pcre_int32 other_case;
1154    } ucd_record;
1155    
1156    extern const ucd_record  _pcre_ucd_records[];
1157    extern const uschar      _pcre_ucd_stage1[];
1158    extern const pcre_uint16 _pcre_ucd_stage2[];
1159    extern const int         _pcre_ucp_gentype[];
1160    
1161    
1162    /* UCD access macros */
1163    
1164    #define UCD_BLOCK_SIZE 128
1165    #define GET_UCD(ch) (_pcre_ucd_records + \
1166            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1167            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1168    
1169    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1170    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1171    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1172    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1173    
1174  #endif  #endif
1175    
1176  /* End of pcre_internal.h */  /* End of pcre_internal.h */

Legend:
Removed from v.172  
changed lines
  Added in v.369

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12