/[pcre]/code/tags/pcre-4.0/internal.h
ViewVC logotype

Diff of /code/tags/pcre-4.0/internal.h

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 25 by nigel, Sat Feb 24 21:38:45 2007 UTC revision 49 by nigel, Sat Feb 24 21:39:33 2007 UTC
# Line 3  Line 3 
3  *************************************************/  *************************************************/
4    
5    
 #define PCRE_VERSION       "2.01 21-Oct-1998"  
   
   
6  /* This is a library of functions to support regular expressions whose syntax  /* This is a library of functions to support regular expressions whose syntax
7  and semantics are as close as possible to those of the Perl 5 language. See  and semantics are as close as possible to those of the Perl 5 language. See
8  the file Tech.Notes for some information on the internals.  the file Tech.Notes for some information on the internals.
9    
10  Written by: Philip Hazel <ph10@cam.ac.uk>  Written by: Philip Hazel <ph10@cam.ac.uk>
11    
12             Copyright (c) 1998 University of Cambridge             Copyright (c) 1997-2000 University of Cambridge
13    
14  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
15  Permission is granted to anyone to use this software for any purpose on any  Permission is granted to anyone to use this software for any purpose on any
# Line 28  restrictions: Line 25  restrictions:
25    
26  3. Altered versions must be plainly marked as such, and must not be  3. Altered versions must be plainly marked as such, and must not be
27     misrepresented as being the original software.     misrepresented as being the original software.
28    
29    4. If PCRE is embedded in any software that is released under the GNU
30       General Purpose Licence (GPL), then the terms of that licence shall
31       supersede any condition above with which it is incompatible.
32  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
33  */  */
34    
35  /* This header contains definitions that are shared between the different  /* This header contains definitions that are shared between the different
36  modules, but which are not relevant to the outside. */  modules, but which are not relevant to the outside. */
37    
38    /* Get the definitions provided by running "configure" */
39    
40    #include "config.h"
41    
42  /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),  /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
43  define a macro for memmove() if USE_BCOPY is defined. */  define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
44    is set. Otherwise, include an emulating function for those systems that have
45    neither (there some non-Unix environments where this is the case). This assumes
46    that all calls to memmove are moving strings upwards in store, which is the
47    case in PCRE. */
48    
49  #ifdef USE_BCOPY  #if ! HAVE_MEMMOVE
50  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
51    #if HAVE_BCOPY
52  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
53    #else
54    void *
55    pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)
56    {
57    int i;
58    dest += n;
59    src += n;
60    for (i = 0; i < n; ++i) *(--dest) =  *(--src);
61    }
62    #define memmove(a, b, c) pcre_memmove(a, b, c)
63    #endif
64  #endif  #endif
65    
66  /* Standard C headers plus the external interface definition */  /* Standard C headers plus the external interface definition */
# Line 63  Standard C system should have one. */ Line 84  Standard C system should have one. */
84    
85  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
86    
87  /* Private options flags start at the most significant end of the two bytes.  /* Private options flags start at the most significant end of the four bytes,
88  The public options defined in pcre.h start at the least significant end. Make  but skip the top bit so we can use ints for convenience without getting tangled
89  sure they don't overlap! */  with negative values. The public options defined in pcre.h start at the least
90    significant end. Make sure they don't overlap, though now that we have expanded
91  #define PCRE_FIRSTSET           0x8000  /* first_char is set */  to four bytes there is plenty of space. */
92  #define PCRE_STARTLINE          0x4000  /* start after \n for multiline */  
93  #define PCRE_INGROUP            0x2000  /* compiling inside a group */  #define PCRE_FIRSTSET      0x40000000  /* first_char is set */
94    #define PCRE_REQCHSET      0x20000000  /* req_char is set */
95    #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */
96    #define PCRE_INGROUP       0x08000000  /* compiling inside a group */
97    #define PCRE_ICHANGED      0x04000000  /* i option changes within regex */
98    
99  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
100    
# Line 80  time, run time or study time, respective Line 105  time, run time or study time, respective
105    
106  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
107    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
108     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY)     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8)
109    
110  #define PUBLIC_EXEC_OPTIONS (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL)  #define PUBLIC_EXEC_OPTIONS \
111      (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY)
112    
113  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
114    
115  /* Magic number to provide a small check against being handed junk. */  /* Magic number to provide a small check against being handed junk. */
116    
117  #define MAGIC_NUMBER  0x50435245   /* 'PCRE' */  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
118    
119  /* Miscellaneous definitions */  /* Miscellaneous definitions */
120    
# Line 178  enum { Line 204  enum {
204    
205    OP_CLASS,          /* Match a character class */    OP_CLASS,          /* Match a character class */
206    OP_REF,            /* Match a back reference */    OP_REF,            /* Match a back reference */
207      OP_RECURSE,        /* Match this pattern recursively */
208    
209    OP_ALT,            /* Start of alternation */    OP_ALT,            /* Start of alternation */
210    OP_KET,            /* End of group that doesn't have an unbounded repeat */    OP_KET,            /* End of group that doesn't have an unbounded repeat */
# Line 244  just to accommodate the POSIX wrapper. * Line 271  just to accommodate the POSIX wrapper. *
271  #define ERR26 "malformed number after (?("  #define ERR26 "malformed number after (?("
272  #define ERR27 "conditional group contains more than two branches"  #define ERR27 "conditional group contains more than two branches"
273  #define ERR28 "assertion expected after (?("  #define ERR28 "assertion expected after (?("
274    #define ERR29 "(?p must be followed by )"
275    #define ERR30 "unknown POSIX class name"
276    #define ERR31 "POSIX collating elements are not supported"
277    #define ERR32 "this version of PCRE is not compiled with PCRE_UTF8 support"
278    #define ERR33 "characters with values > 255 are not yet supported in classes"
279    #define ERR34 "character value in \\x{...} sequence is too large"
280    
281  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
282  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
# Line 258  typedef unsigned char uschar; Line 291  typedef unsigned char uschar;
291  runs on as long as necessary after the end. */  runs on as long as necessary after the end. */
292    
293  typedef struct real_pcre {  typedef struct real_pcre {
294    unsigned int  magic_number;    unsigned long int magic_number;
295      size_t size;
296    const unsigned char *tables;    const unsigned char *tables;
297    unsigned short int options;    unsigned long int options;
298    unsigned char top_bracket;    uschar top_bracket;
299    unsigned char top_backref;    uschar top_backref;
300    unsigned char first_char;    uschar first_char;
301    unsigned char code[1];    uschar req_char;
302      uschar code[1];
303  } real_pcre;  } real_pcre;
304    
305  /* The real format of the extra block returned by pcre_study(). */  /* The real format of the extra block returned by pcre_study(). */
306    
307  typedef struct real_pcre_extra {  typedef struct real_pcre_extra {
308    unsigned char options;    uschar options;
309    unsigned char start_bits[32];    uschar start_bits[32];
310  } real_pcre_extra;  } real_pcre_extra;
311    
312    
# Line 280  doing the compiling, so that they are th Line 315  doing the compiling, so that they are th
315    
316  typedef struct compile_data {  typedef struct compile_data {
317    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
318    const uschar *fcc;            /* Points to case-flippint table */    const uschar *fcc;            /* Points to case-flipping table */
319    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
320    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
321  } compile_data;  } compile_data;
# Line 298  typedef struct match_data { Line 333  typedef struct match_data {
333    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
334    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
335    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
336      BOOL   utf8;                  /* UTF8 flag */
337    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
338      BOOL   notempty;              /* Empty string match not wanted */
339      const uschar *start_pattern;  /* For use when recursing */
340    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
341    const uschar *end_subject;    /* End of the subject string */    const uschar *end_subject;    /* End of the subject string */
342      const uschar *start_match;    /* Start of this match attempt */
343    const uschar *end_match_ptr;  /* Subject position at end match */    const uschar *end_match_ptr;  /* Subject position at end match */
344    int     end_offset_top;       /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
345  } match_data;  } match_data;
346    
347  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 315  typedef struct match_data { Line 354  typedef struct match_data {
354  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
355    
356  /* 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
357  of bits for a class map. */  of bits for a class map. Some classes are built by combining these tables. */
358    
359  #define cbit_digit    0      /* for \d */  #define cbit_space     0      /* [:space:] or \s */
360  #define cbit_word    32      /* for \w */  #define cbit_xdigit   32      /* [:xdigit:] */
361  #define cbit_space   64      /* for \s */  #define cbit_digit    64      /* [:digit:] or \d */
362  #define cbit_length  96      /* Length of the cbits table */  #define cbit_upper    96      /* [:upper:] */
363    #define cbit_lower   128      /* [:lower:] */
364    #define cbit_word    160      /* [:word:] or \w */
365    #define cbit_graph   192      /* [:graph:] */
366    #define cbit_print   224      /* [:print:] */
367    #define cbit_punct   256      /* [:punct:] */
368    #define cbit_cntrl   288      /* [:cntrl:] */
369    #define cbit_length  320      /* Length of the cbits table */
370    
371  /* Offsets of the various tables from the base tables pointer, and  /* Offsets of the various tables from the base tables pointer, and
372  total length. */  total length. */
Legend:
Removed from v.25  
changed lines
  Added in v.49
  ViewVC Help
Powered by ViewVC 1.1.5