/[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 199 by ph10, Tue Jul 31 14:39:09 2007 UTC revision 335 by ph10, Sat Apr 12 14:36:14 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 358  capturing parenthesis numbers in back re Line 358  capturing parenthesis numbers in back re
358    
359  /* When UTF-8 encoding is being used, a character is no longer just a single  /* When UTF-8 encoding is being used, a character is no longer just a single
360  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
361  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
362    never be called in byte mode. To make sure it can never even appear when UTF-8
363    support is omitted, we don't even define it. */
364    
365  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
366    #define NEXTCHAR(p) p++;
367  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
368  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
369  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
370  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
371  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
372  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
373    
374  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
375    
376    /* Advance a character pointer one byte in non-UTF-8 mode and by one character
377    in UTF-8 mode. */
378    
379    #define NEXTCHAR(p) \
380      p++; \
381      if (utf8) { while((*p & 0xc0) == 0x80) p++; }
382    
383  /* Get the next UTF-8 character, not advancing the pointer. This is called when  /* Get the next UTF-8 character, not advancing the pointer. This is called when
384  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
385    
# Line 459  if there are extra bytes. This is called Line 469  if there are extra bytes. This is called
469      }      }
470    
471  /* If the pointer is not at the start of a character, move it back until  /* If the pointer is not at the start of a character, move it back until
472  it is. Called only in UTF-8 mode. */  it is. This is called only in UTF-8 mode - we don't put a test within the macro
473    because almost all calls are already within a block of UTF-8 only code. */
474    
475  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
476    
477  #endif  #endif
478    
# Line 478  Standard C system should have one. */ Line 489  Standard C system should have one. */
489    
490  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
491    
492  /* Private options flags start at the most significant end of the four bytes.  /* Private flags containing information about the compiled regex. They used to
493  The public options defined in pcre.h start at the least significant end. Make  live at the top end of the options word, but that got almost full, so now they
494  sure they don't overlap! The bits are getting a bit scarce now -- when we run  are in a 16-bit flags word. */
495  out, there is a dummy word in the structure that could be used for the private  
496  bits. */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
497    #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
498  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
499  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
500  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
501  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
 #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */  
502    
503  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
504    
# Line 504  time, run time, or study time, respectiv Line 514  time, run time, or study time, respectiv
514    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
515     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
516     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
517     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
518    
519  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
520    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
521     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
522    
523  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
524    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
525     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
526       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
527    
528  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
529    
# Line 602  character, that code will have to change Line 613  character, that code will have to change
613    
614  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,
615         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_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,
616         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_k, ESC_REF };         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k, ESC_REF };
617    
618    
619  /* 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  
620  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.
621    
 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.  
   
622  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
623  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
624  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" in pcre_dfa_exec.c that must be updated. */
# Line 755  enum { Line 761  enum {
761    OP_DEF,            /* 101 The DEFINE condition */    OP_DEF,            /* 101 The DEFINE condition */
762    
763    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_BRAZERO,        /* 102 These two must remain together and in this */
764    OP_BRAMINZERO      /* 103 order. */    OP_BRAMINZERO,     /* 103 order. */
765    
766      /* These are backtracking control verbs */
767    
768      OP_PRUNE,          /* 104 */
769      OP_SKIP,           /* 105 */
770      OP_THEN,           /* 106 */
771      OP_COMMIT,         /* 107 */
772    
773      /* These are forced failure and success verbs */
774    
775      OP_FAIL,           /* 108 */
776      OP_ACCEPT,         /* 109 */
777    
778      /* This is used to skip a subpattern with a {0} quantifier */
779    
780      OP_SKIPZERO        /* 110 */
781  };  };
782    
783    
# Line 779  for debugging. The macro is referenced o Line 801  for debugging. The macro is referenced o
801    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
802    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
803    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
804    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
805      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
806      "Skip zero"
807    
808    
809  /* 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 843  in UTF-8 mode. The code that uses this t Line 867  in UTF-8 mode. The code that uses this t
867    3,                             /* RREF                                   */ \    3,                             /* RREF                                   */ \
868    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
869    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
870      1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
871      1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
872    
873    
874  /* 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 883  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
883         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
884         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
885         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
886         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
887           ERR60, ERR61, ERR62, ERR63 };
888    
889  /* 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
890  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 906  NOTE NOTE NOTE:
906  typedef struct real_pcre {  typedef struct real_pcre {
907    pcre_uint32 magic_number;    pcre_uint32 magic_number;
908    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
909    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
910    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
911      pcre_uint16 dummy1;             /* For future use */
912    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
913    pcre_uint16 top_backref;    pcre_uint16 top_backref;
914    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 920  typedef struct compile_data { Line 947  typedef struct compile_data {
947    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
948    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
949    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
950    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
951      int  final_bracount;          /* Saved value after first pass */
952    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
953    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
954    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
955      int  external_flags;          /* External flag bits to be set */
956    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
957    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
958    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
959    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
960    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 984  typedef struct match_data { Line 1013  typedef struct match_data {
1013    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1014    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1015    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1016      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1017    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1018    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1019    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
# Line 1020  typedef struct dfa_match_data { Line 1050  typedef struct dfa_match_data {
1050  #define ctype_letter  0x02  #define ctype_letter  0x02
1051  #define ctype_digit   0x04  #define ctype_digit   0x04
1052  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1053  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1054  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1055    
1056  /* 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 1078  total length. */
1078  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1079    
1080  /* 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
1081  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1082    relocations in shared libraries, it now has an offset into a single string
1083    instead. */
1084    
1085  typedef struct {  typedef struct {
1086    const char *name;    pcre_uint16 name_offset;
1087    pcre_uint16 type;    pcre_uint16 type;
1088    pcre_uint16 value;    pcre_uint16 value;
1089  } ucp_type_table;  } ucp_type_table;
# Line 1069  extern const uschar _pcre_utf8_table4[]; Line 1101  extern const uschar _pcre_utf8_table4[];
1101    
1102  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1103    
1104    extern const char   _pcre_utt_names[];
1105  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1106  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1107    
Legend:
Removed from v.199  
changed lines
  Added in v.335
webmaster@exim.org
 ViewVC Help
Powered by ViewVC 1.1.12