/[pcre]/code/trunk/internal.h
ViewVC logotype

Diff of /code/trunk/internal.h

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

revision 9 by nigel, Sat Feb 24 21:38:13 2007 UTC revision 41 by nigel, Sat Feb 24 21:39:17 2007 UTC
# Line 3  Line 3 
3  *************************************************/  *************************************************/
4    
5    
 #define PCRE_VERSION       "1.03 18-Dec-1997"  
   
   
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) 1997 University of Cambridge             Copyright (c) 1997-1999 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 not defined. */
44    
45  #ifdef USE_BCOPY  #ifndef HAVE_MEMMOVE
46  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
47  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
48  #endif  #endif
# Line 46  define a macro for memmove() if USE_BCOP Line 51  define a macro for memmove() if USE_BCOP
51    
52  #include <ctype.h>  #include <ctype.h>
53  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
54  #include <stddef.h>  #include <stddef.h>
55  #include <stdio.h>  #include <stdio.h>
56  #include <stdlib.h>  #include <stdlib.h>
# Line 57  define a macro for memmove() if USE_BCOP Line 61  define a macro for memmove() if USE_BCOP
61  Standard C system should have one. */  Standard C system should have one. */
62    
63  #ifndef offsetof  #ifndef offsetof
64  #define offsetof(p_type,field) ((size_t)&(((p_type)0)->field))  #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
65  #endif  #endif
66    
67  /* Private options flags start at the most significant end of the two bytes.  /* These are the public options that can change during matching. */
68  The public options defined in pcre.h start at the least significant end. Make  
69  sure they don't overlap! */  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
70    
71  #define PCRE_FIRSTSET           0x8000  /* first_char is set */  /* Private options flags start at the most significant end of the four bytes,
72  #define PCRE_STARTLINE          0x4000  /* start after \n for multiline */  but skip the top bit so we can use ints for convenience without getting tangled
73  #define PCRE_COMPILED_CASELESS  0x2000  /* like it says */  with negative values. The public options defined in pcre.h start at the least
74    significant end. Make sure they don't overlap, though now that we have expanded
75    to four bytes there is plenty of space. */
76    
77    #define PCRE_FIRSTSET      0x40000000  /* first_char is set */
78    #define PCRE_REQCHSET      0x20000000  /* req_char is set */
79    #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */
80    #define PCRE_INGROUP       0x08000000  /* compiling inside a group */
81    #define PCRE_ICHANGED      0x04000000  /* i option changes within regex */
82    
83  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
84    
85  #define PCRE_STUDY_CASELESS 0x01     /* study was caseless */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
 #define PCRE_STUDY_MAPPED   0x02     /* a map of starting chars exists */  
86    
87  /* Masks for identifying the public options: all permitted at compile time,  /* Masks for identifying the public options which are permitted at compile
88  only some permitted at run or study time. */  time, run time or study time, respectively. */
89    
90  #define PUBLIC_OPTIONS \  #define PUBLIC_OPTIONS \
91    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
92     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA)     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY)
93    
94  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
95    (PCRE_CASELESS|PCRE_ANCHORED|PCRE_MULTILINE|PCRE_NOTBOL|PCRE_NOTEOL| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY)
    PCRE_DOTALL|PCRE_DOLLAR_ENDONLY)  
96    
97  #define PUBLIC_STUDY_OPTIONS (PCRE_CASELESS)  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
98    
99  /* Magic number to provide a small check against being handed junk. */  /* Magic number to provide a small check against being handed junk. */
100    
101  #define MAGIC_NUMBER  0x50435245   /* 'PCRE' */  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
102    
103  /* Miscellaneous definitions */  /* Miscellaneous definitions */
104    
# Line 100  typedef int BOOL; Line 110  typedef int BOOL;
110  /* 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
111  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
112  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
113  definitions below, up to ESC_Z. The final one must be ESC_REF as subsequent  definitions below, up to ESC_z. The final one must be ESC_REF as subsequent
114  values are used for \1, \2, \3, etc. There is a test in the code for an escape  values are used for \1, \2, \3, etc. There is a test in the code for an escape
115  greater than ESC_b and less than ESC_X to detect the types that may be  greater than ESC_b and less than ESC_X to detect the types that may be
116  repeated. If any new escapes are put in-between that don't consume a character,  repeated. If any new escapes are put in-between that don't consume a character,
117  that code will have to change. */  that code will have to change. */
118    
119  enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,  enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,
120           ESC_Z, ESC_z, ESC_REF };
                     /* These are not Perl escapes, so can't appear in the */  
        ESC_X,       /* simple table-lookup because they must be conditional */  
                     /* on PCRE_EXTRA. */  
        ESC_Z,  
        ESC_REF };  
121    
122  /* 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
123  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
# Line 132  enum { Line 137  enum {
137    OP_WHITESPACE,         /* \s */    OP_WHITESPACE,         /* \s */
138    OP_NOT_WORDCHAR,       /* \W */    OP_NOT_WORDCHAR,       /* \W */
139    OP_WORDCHAR,           /* \w */    OP_WORDCHAR,           /* \w */
140    OP_CUT,            /* The analogue of Prolog's "cut" operation (extension) */    OP_EODN,           /* End of data or \n at end of data: \Z. */
141    OP_EOD,            /* End of data: \Z. */    OP_EOD,            /* End of data: \z */
142    
143      OP_OPT,            /* Set runtime options */
144    OP_CIRC,           /* Start of line - varies with multiline switch */    OP_CIRC,           /* Start of line - varies with multiline switch */
145    OP_DOLL,           /* End of line - varies with multiline switch */    OP_DOLL,           /* End of line - varies with multiline switch */
146    OP_ANY,            /* Match any character */    OP_ANY,            /* Match any character */
# Line 188  enum { Line 194  enum {
194    OP_KETRMAX,        /* These two must remain together and in this */    OP_KETRMAX,        /* These two must remain together and in this */
195    OP_KETRMIN,        /* order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* order. They are for groups the repeat for ever. */
196    
197    OP_ASSERT,    /* The assertions must come before ONCE and COND */
198    OP_ASSERT_NOT,  
199      OP_ASSERT,         /* Positive lookahead */
200      OP_ASSERT_NOT,     /* Negative lookahead */
201      OP_ASSERTBACK,     /* Positive lookbehind */
202      OP_ASSERTBACK_NOT, /* Negative lookbehind */
203      OP_REVERSE,        /* Move pointer back - used in lookbehind assertions */
204    
205      /* ONCE and COND must come after the assertions, with ONCE first, as there's
206      a test for >= ONCE for a subpattern that isn't an assertion. */
207    
208    OP_ONCE,           /* Once matched, don't back up into the subpattern */    OP_ONCE,           /* Once matched, don't back up into the subpattern */
209      OP_COND,           /* Conditional group */
210      OP_CREF,           /* Used to hold an extraction string number */
211    
212    OP_BRAZERO,        /* These two must remain together and in this */    OP_BRAZERO,        /* These two must remain together and in this */
213    OP_BRAMINZERO,     /* order. */    OP_BRAMINZERO,     /* order. */
# Line 230  just to accommodate the POSIX wrapper. * Line 247  just to accommodate the POSIX wrapper. *
247  #define ERR19 "too many sets of parentheses"  #define ERR19 "too many sets of parentheses"
248  #define ERR20 "regular expression too large"  #define ERR20 "regular expression too large"
249  #define ERR21 "failed to get memory"  #define ERR21 "failed to get memory"
250  #define ERR22 "unmatched brackets"  #define ERR22 "unmatched parentheses"
251  #define ERR23 "internal error: code overflow"  #define ERR23 "internal error: code overflow"
252    #define ERR24 "unrecognized character after (?<"
253    #define ERR25 "lookbehind assertion is not fixed length"
254    #define ERR26 "malformed number after (?("
255    #define ERR27 "conditional group contains more than two branches"
256    #define ERR28 "assertion expected after (?("
257    
258  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
259  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 246  typedef unsigned char uschar; Line 268  typedef unsigned char uschar;
268  runs on as long as necessary after the end. */  runs on as long as necessary after the end. */
269    
270  typedef struct real_pcre {  typedef struct real_pcre {
271    unsigned int  magic_number;    unsigned long int magic_number;
272    unsigned short int options;    const unsigned char *tables;
273    unsigned char top_bracket;    unsigned long int options;
274    unsigned char top_backref;    uschar top_bracket;
275    unsigned char first_char;    uschar top_backref;
276    unsigned char code[1];    uschar first_char;
277      uschar req_char;
278      uschar code[1];
279  } real_pcre;  } real_pcre;
280    
281  /* The real format of the extra block returned by pcre_study(). */  /* The real format of the extra block returned by pcre_study(). */
282    
283  typedef struct real_pcre_extra {  typedef struct real_pcre_extra {
284    unsigned char options;    uschar options;
285    unsigned char start_bits[32];    uschar start_bits[32];
286  } real_pcre_extra;  } real_pcre_extra;
287    
 /* Global tables from chartables.c */  
288    
289  extern uschar pcre_lcc[];  /* Structure for passing "static" information around between the functions
290  extern uschar pcre_fcc[];  doing the compiling, so that they are thread-safe. */
291  extern uschar pcre_cbits[];  
292  extern uschar pcre_ctypes[];  typedef struct compile_data {
293      const uschar *lcc;            /* Points to lower casing table */
294      const uschar *fcc;            /* Points to case-flipping table */
295      const uschar *cbits;          /* Points to character type table */
296      const uschar *ctypes;         /* Points to table of type maps */
297    } compile_data;
298    
299    /* Structure for passing "static" information around between the functions
300    doing the matching, so that they are thread-safe. */
301    
302    typedef struct match_data {
303      int    errorcode;             /* As it says */
304      int   *offset_vector;         /* Offset vector */
305      int    offset_end;            /* One past the end */
306      int    offset_max;            /* The maximum usable for return data */
307      const uschar *lcc;            /* Points to lower casing table */
308      const uschar *ctypes;         /* Points to table of type maps */
309      BOOL   offset_overflow;       /* Set if too many extractions */
310      BOOL   notbol;                /* NOTBOL flag */
311      BOOL   noteol;                /* NOTEOL flag */
312      BOOL   endonly;               /* Dollar not before final \n */
313      BOOL   notempty;              /* Empty string match not wanted */
314      const uschar *start_subject;  /* Start of the subject string */
315      const uschar *end_subject;    /* End of the subject string */
316      const uschar *start_match;    /* Start of this match attempt */
317      const uschar *end_match_ptr;  /* Subject position at end match */
318      int     end_offset_top;       /* Highwater mark at end of match */
319    } match_data;
320    
321  /* Bit definitions for entries in pcre_ctypes[]. */  /* Bit definitions for entries in the pcre_ctypes table. */
322    
323  #define ctype_space   0x01  #define ctype_space   0x01
324  #define ctype_letter  0x02  #define ctype_letter  0x02
# Line 277  extern uschar pcre_ctypes[]; Line 327  extern uschar pcre_ctypes[];
327  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphameric or '_' */
328  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
329    
330  /* Offsets for the bitmap tables */  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
331    of bits for a class map. */
332    
333  #define cbit_digit    0  #define cbit_digit    0      /* for \d */
334  #define cbit_letter  32  #define cbit_word    32      /* for \w */
335  #define cbit_word    64  #define cbit_space   64      /* for \s */
336  #define cbit_space   96  #define cbit_length  96      /* Length of the cbits table */
337  #define cbit_length 128      /* Length of the cbits table */  
338    /* Offsets of the various tables from the base tables pointer, and
339    total length. */
340    
341    #define lcc_offset      0
342    #define fcc_offset    256
343    #define cbits_offset  512
344    #define ctypes_offset (cbits_offset + cbit_length)
345    #define tables_length (ctypes_offset + 256)
346    
347  /* End of internal.h */  /* End of internal.h */

Legend:
Removed from v.9  
changed lines
  Added in v.41

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12