/[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 85 by nigel, Sat Feb 24 21:41:13 2007 UTC revision 87 by nigel, Sat Feb 24 21:41:21 2007 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) 1997-2005 University of Cambridge             Copyright (c) 1997-2006 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 114  Unix, where it is defined in sys/types, Line 114  Unix, where it is defined in sys/types,
114    
115  typedef unsigned char uschar;  typedef unsigned char uschar;
116    
117    /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
118    with a custom type. This makes it possible, for example, to allow pcre_exec()
119    to process subject strings that are discontinuous by using a smart pointer
120    class. It must always be possible to inspect all of the subject string in
121    pcre_exec() because of the way it backtracks. Two macros are required in the
122    normal case, for sign-unspecified and unsigned char pointers. The former is
123    used for the external interface and appears in pcre.h, which is why its name
124    must begin with PCRE_. */
125    
126    #ifdef CUSTOM_SUBJECT_PTR
127    #define PCRE_SPTR CUSTOM_SUBJECT_PTR
128    #define USPTR CUSTOM_SUBJECT_PTR
129    #else
130    #define PCRE_SPTR const char *
131    #define USPTR const unsigned char *
132    #endif
133    
134  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
135  values. */  values. */
136    
# Line 146  case in PCRE. */ Line 163  case in PCRE. */
163  void *  void *
164  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)
165  {  {
166  int i;  size_t i;
167  dest += n;  dest += n;
168  src += n;  src += n;
169  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  for (i = 0; i < n; ++i) *(--dest) =  *(--src);
170    return dest;
171  }  }
172  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
173  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 443  ESC_n is defined as yet another macro, w Line 461  ESC_n is defined as yet another macro, w
461  #define ESC_tee '\t'  #define ESC_tee '\t'
462  #endif  #endif
463    
464    /* Codes for different types of Unicode property */
465    
466    #define PT_ANY        0    /* Any property - matches all chars */
467    #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
468    #define PT_GC         2    /* General characteristic (e.g. L) */
469    #define PT_PC         3    /* Particular characteristic (e.g. Lu) */
470    #define PT_SC         4    /* Script (e.g. Han) */
471    
472    /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
473    contain UTF-8 characters with values greater than 255. */
474    
475    #define XCL_NOT    0x01    /* Flag: this is a negative class */
476    #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */
477    
478    #define XCL_END       0    /* Marks end of individual items */
479    #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */
480    #define XCL_RANGE     2    /* A range (two multibyte chars) follows */
481    #define XCL_PROP      3    /* Unicode property (2-byte property code follows) */
482    #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
483    
484  /* These are escaped items that aren't just an encoding of a particular data  /* These are escaped items that aren't just an encoding of a particular data
485  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
486  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
# Line 458  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, E Line 496  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, E
496         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E,         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E,
497         ESC_Q, ESC_REF };         ESC_Q, ESC_REF };
498    
 /* Flag bits and data types for the extended class (OP_XCLASS) for classes that  
 contain UTF-8 characters with values greater than 255. */  
   
 #define XCL_NOT    0x01    /* Flag: this is a negative class */  
 #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */  
   
 #define XCL_END       0    /* Marks end of individual items */  
 #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */  
 #define XCL_RANGE     2    /* A range (two multibyte chars) follows */  
 #define XCL_PROP      3    /* Unicode property (one property code) follows */  
 #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  
   
   
499  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
500  that extract substrings. 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
501  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.
# Line 634  in UTF-8 mode. The code that uses this t Line 659  in UTF-8 mode. The code that uses this t
659    1,                             /* End                                    */ \    1,                             /* End                                    */ \
660    1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \    1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \
661    1, 1,                          /* Any, Anybyte                           */ \    1, 1,                          /* Any, Anybyte                           */ \
662    2, 2, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
663    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
664    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
665    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 766  typedef struct recursion_info { Line 791  typedef struct recursion_info {
791    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
792    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
793    const uschar *after_call;     /* "Return value": points after the call in the expr */    const uschar *after_call;     /* "Return value": points after the call in the expr */
794    const uschar *save_start;     /* Old value of md->start_match */    USPTR save_start;             /* Old value of md->start_match */
795    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
796    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
797  } recursion_info;  } recursion_info;
# Line 785  struct heapframe; Line 810  struct heapframe;
810  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
811    
812  typedef struct match_data {  typedef struct match_data {
813    unsigned long int match_call_count; /* As it says */    unsigned long int match_call_count;      /* As it says */
814    unsigned long int match_limit;/* As it says */    unsigned long int match_limit;           /* As it says */
815      unsigned long int match_limit_recursion; /* As it says */
816    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
817    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
818    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
# Line 801  typedef struct match_data { Line 827  typedef struct match_data {
827    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
828    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
829    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
830    const uschar *start_subject;  /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
831    const uschar *end_subject;    /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
832    const uschar *start_match;    /* Start of this match attempt */    USPTR  start_match;           /* Start of this match attempt */
833    const uschar *end_match_ptr;  /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
834    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
835    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
836    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 859  total length. */ Line 885  total length. */
885  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
886  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
887    
888  /* Layout of the UCP type table that translates property names into codes for  /* Layout of the UCP type table that translates property names into types and
889  pcre_ucp_findchar(). */  codes. */
890    
891  typedef struct {  typedef struct {
892    const char *name;    const char *name;
893    int value;    pcre_uint16 type;
894      pcre_uint16 value;
895  } ucp_type_table;  } ucp_type_table;
896    
897    
# Line 895  sense, but are not part of the PCRE publ Line 922  sense, but are not part of the PCRE publ
922  extern int         _pcre_ord2utf8(int, uschar *);  extern int         _pcre_ord2utf8(int, uschar *);
923  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,
924                       const pcre_study_data *, pcre_study_data *);                       const pcre_study_data *, pcre_study_data *);
925  extern int         _pcre_ucp_findchar(const int, int *, int *);  extern int         _pcre_ucp_findprop(const int, int *, int *);
926    extern int         _pcre_ucp_othercase(const int);
927  extern int         _pcre_valid_utf8(const uschar *, int);  extern int         _pcre_valid_utf8(const uschar *, int);
928  extern BOOL        _pcre_xclass(int, const uschar *);  extern BOOL        _pcre_xclass(int, const uschar *);
929    
Legend:
Removed from v.85  
changed lines
  Added in v.87
  ViewVC Help
Powered by ViewVC 1.1.5