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

Contents of /code/trunk/internal.h

Parent Directory Parent Directory | Revision Log Revision Log


Revision 31 - (hide annotations) (download)
Sat Feb 24 21:38:57 2007 UTC (7 years, 7 months ago) by nigel
File MIME type: text/plain
File size: 13536 byte(s)
Load pcre-2.04 into code/trunk.

1 nigel 3 /*************************************************
2     * Perl-Compatible Regular Expressions *
3     *************************************************/
4    
5    
6 nigel 31 #define PCRE_VERSION "2.04 19-Feb-1999"
7 nigel 3
8    
9     /* This is a library of functions to support regular expressions whose syntax
10     and semantics are as close as possible to those of the Perl 5 language. See
11     the file Tech.Notes for some information on the internals.
12    
13     Written by: Philip Hazel <ph10@cam.ac.uk>
14    
15 nigel 27 Copyright (c) 1997-1999 University of Cambridge
16 nigel 3
17     -----------------------------------------------------------------------------
18     Permission is granted to anyone to use this software for any purpose on any
19     computer system, and to redistribute it freely, subject to the following
20     restrictions:
21    
22     1. This software is distributed in the hope that it will be useful,
23     but WITHOUT ANY WARRANTY; without even the implied warranty of
24     MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
25    
26     2. The origin of this software must not be misrepresented, either by
27     explicit claim or by omission.
28    
29     3. Altered versions must be plainly marked as such, and must not be
30     misrepresented as being the original software.
31 nigel 29
32     4. If PCRE is embedded in any software that is released under the GNU
33     General Purpose Licence (GPL), then the terms of that licence shall
34     supersede any condition above with which it is incompatible.
35 nigel 3 -----------------------------------------------------------------------------
36     */
37    
38     /* This header contains definitions that are shared between the different
39     modules, but which are not relevant to the outside. */
40    
41     /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
42     define a macro for memmove() if USE_BCOPY is defined. */
43    
44     #ifdef USE_BCOPY
45 nigel 9 #undef memmove /* some systems may have a macro */
46 nigel 3 #define memmove(a, b, c) bcopy(b, a, c)
47     #endif
48    
49     /* Standard C headers plus the external interface definition */
50    
51     #include <ctype.h>
52     #include <limits.h>
53     #include <stddef.h>
54     #include <stdio.h>
55     #include <stdlib.h>
56     #include <string.h>
57     #include "pcre.h"
58    
59 nigel 9 /* In case there is no definition of offsetof() provided - though any proper
60     Standard C system should have one. */
61    
62     #ifndef offsetof
63 nigel 11 #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
64 nigel 9 #endif
65    
66 nigel 23 /* These are the public options that can change during matching. */
67    
68     #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
69    
70 nigel 3 /* Private options flags start at the most significant end of the two bytes.
71     The public options defined in pcre.h start at the least significant end. Make
72     sure they don't overlap! */
73    
74     #define PCRE_FIRSTSET 0x8000 /* first_char is set */
75     #define PCRE_STARTLINE 0x4000 /* start after \n for multiline */
76 nigel 23 #define PCRE_INGROUP 0x2000 /* compiling inside a group */
77 nigel 3
78     /* Options for the "extra" block produced by pcre_study(). */
79    
80 nigel 23 #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
81 nigel 3
82 nigel 23 /* Masks for identifying the public options which are permitted at compile
83     time, run time or study time, respectively. */
84 nigel 3
85     #define PUBLIC_OPTIONS \
86     (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
87 nigel 19 PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY)
88 nigel 3
89 nigel 23 #define PUBLIC_EXEC_OPTIONS (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL)
90 nigel 3
91 nigel 23 #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
92 nigel 3
93     /* Magic number to provide a small check against being handed junk. */
94    
95     #define MAGIC_NUMBER 0x50435245 /* 'PCRE' */
96    
97     /* Miscellaneous definitions */
98    
99     typedef int BOOL;
100    
101     #define FALSE 0
102     #define TRUE 1
103    
104     /* These are escaped items that aren't just an encoding of a particular data
105     value such as \n. They must have non-zero values, as check_escape() returns
106     their negation. Also, they must appear in the same order as in the opcode
107 nigel 23 definitions below, up to ESC_z. The final one must be ESC_REF as subsequent
108 nigel 3 values are used for \1, \2, \3, etc. There is a test in the code for an escape
109     greater than ESC_b and less than ESC_X to detect the types that may be
110     repeated. If any new escapes are put in-between that don't consume a character,
111     that code will have to change. */
112    
113     enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,
114 nigel 23 ESC_Z, ESC_z, ESC_REF };
115 nigel 3
116     /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
117     that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
118 nigel 7 OP_EOD must correspond in order to the list of escapes immediately above. */
119 nigel 3
120     enum {
121     OP_END, /* End of pattern */
122    
123     /* Values corresponding to backslashed metacharacters */
124    
125     OP_SOD, /* Start of data: \A */
126     OP_NOT_WORD_BOUNDARY, /* \B */
127     OP_WORD_BOUNDARY, /* \b */
128     OP_NOT_DIGIT, /* \D */
129     OP_DIGIT, /* \d */
130     OP_NOT_WHITESPACE, /* \S */
131     OP_WHITESPACE, /* \s */
132     OP_NOT_WORDCHAR, /* \W */
133     OP_WORDCHAR, /* \w */
134 nigel 23 OP_EODN, /* End of data or \n at end of data: \Z. */
135     OP_EOD, /* End of data: \z */
136 nigel 3
137 nigel 23 OP_OPT, /* Set runtime options */
138 nigel 3 OP_CIRC, /* Start of line - varies with multiline switch */
139     OP_DOLL, /* End of line - varies with multiline switch */
140     OP_ANY, /* Match any character */
141     OP_CHARS, /* Match string of characters */
142     OP_NOT, /* Match anything but the following char */
143    
144     OP_STAR, /* The maximizing and minimizing versions of */
145     OP_MINSTAR, /* all these opcodes must come in pairs, with */
146     OP_PLUS, /* the minimizing one second. */
147     OP_MINPLUS, /* This first set applies to single characters */
148     OP_QUERY,
149     OP_MINQUERY,
150     OP_UPTO, /* From 0 to n matches */
151     OP_MINUPTO,
152     OP_EXACT, /* Exactly n matches */
153    
154     OP_NOTSTAR, /* The maximizing and minimizing versions of */
155     OP_NOTMINSTAR, /* all these opcodes must come in pairs, with */
156     OP_NOTPLUS, /* the minimizing one second. */
157     OP_NOTMINPLUS, /* This first set applies to "not" single characters */
158     OP_NOTQUERY,
159     OP_NOTMINQUERY,
160     OP_NOTUPTO, /* From 0 to n matches */
161     OP_NOTMINUPTO,
162     OP_NOTEXACT, /* Exactly n matches */
163    
164     OP_TYPESTAR, /* The maximizing and minimizing versions of */
165     OP_TYPEMINSTAR, /* all these opcodes must come in pairs, with */
166     OP_TYPEPLUS, /* the minimizing one second. These codes must */
167     OP_TYPEMINPLUS, /* be in exactly the same order as those above. */
168     OP_TYPEQUERY, /* This set applies to character types such as \d */
169     OP_TYPEMINQUERY,
170     OP_TYPEUPTO, /* From 0 to n matches */
171     OP_TYPEMINUPTO,
172     OP_TYPEEXACT, /* Exactly n matches */
173    
174     OP_CRSTAR, /* The maximizing and minimizing versions of */
175     OP_CRMINSTAR, /* all these opcodes must come in pairs, with */
176     OP_CRPLUS, /* the minimizing one second. These codes must */
177     OP_CRMINPLUS, /* be in exactly the same order as those above. */
178     OP_CRQUERY, /* These are for character classes and back refs */
179     OP_CRMINQUERY,
180     OP_CRRANGE, /* These are different to the three seta above. */
181     OP_CRMINRANGE,
182    
183     OP_CLASS, /* Match a character class */
184     OP_REF, /* Match a back reference */
185    
186     OP_ALT, /* Start of alternation */
187     OP_KET, /* End of group that doesn't have an unbounded repeat */
188     OP_KETRMAX, /* These two must remain together and in this */
189     OP_KETRMIN, /* order. They are for groups the repeat for ever. */
190    
191 nigel 23 /* The assertions must come before ONCE and COND */
192    
193     OP_ASSERT, /* Positive lookahead */
194     OP_ASSERT_NOT, /* Negative lookahead */
195     OP_ASSERTBACK, /* Positive lookbehind */
196     OP_ASSERTBACK_NOT, /* Negative lookbehind */
197     OP_REVERSE, /* Move pointer back - used in lookbehind assertions */
198    
199     /* ONCE and COND must come after the assertions, with ONCE first, as there's
200     a test for >= ONCE for a subpattern that isn't an assertion. */
201    
202 nigel 3 OP_ONCE, /* Once matched, don't back up into the subpattern */
203 nigel 23 OP_COND, /* Conditional group */
204     OP_CREF, /* Used to hold an extraction string number */
205 nigel 3
206     OP_BRAZERO, /* These two must remain together and in this */
207     OP_BRAMINZERO, /* order. */
208    
209     OP_BRA /* This and greater values are used for brackets that
210     extract substrings. */
211     };
212    
213     /* The highest extraction number. This is limited by the number of opcodes
214     left after OP_BRA, i.e. 255 - OP_BRA. We actually set it somewhat lower. */
215    
216     #define EXTRACT_MAX 99
217    
218     /* The texts of compile-time error messages are defined as macros here so that
219     they can be accessed by the POSIX wrapper and converted into error codes. Yes,
220     I could have used error codes in the first place, but didn't feel like changing
221     just to accommodate the POSIX wrapper. */
222    
223     #define ERR1 "\\ at end of pattern"
224     #define ERR2 "\\c at end of pattern"
225     #define ERR3 "unrecognized character follows \\"
226     #define ERR4 "numbers out of order in {} quantifier"
227     #define ERR5 "number too big in {} quantifier"
228     #define ERR6 "missing terminating ] for character class"
229     #define ERR7 "invalid escape sequence in character class"
230     #define ERR8 "range out of order in character class"
231     #define ERR9 "nothing to repeat"
232     #define ERR10 "operand of unlimited repeat could match the empty string"
233     #define ERR11 "internal error: unexpected repeat"
234     #define ERR12 "unrecognized character after (?"
235     #define ERR13 "too many capturing parenthesized sub-patterns"
236     #define ERR14 "missing )"
237     #define ERR15 "back reference to non-existent subpattern"
238     #define ERR16 "erroffset passed as NULL"
239     #define ERR17 "unknown option bit(s) set"
240     #define ERR18 "missing ) after comment"
241     #define ERR19 "too many sets of parentheses"
242     #define ERR20 "regular expression too large"
243     #define ERR21 "failed to get memory"
244 nigel 23 #define ERR22 "unmatched parentheses"
245 nigel 3 #define ERR23 "internal error: code overflow"
246 nigel 23 #define ERR24 "unrecognized character after (?<"
247     #define ERR25 "lookbehind assertion is not fixed length"
248     #define ERR26 "malformed number after (?("
249     #define ERR27 "conditional group contains more than two branches"
250     #define ERR28 "assertion expected after (?("
251 nigel 3
252     /* All character handling must be done as unsigned characters. Otherwise there
253     are problems with top-bit-set characters and functions such as isspace().
254     However, we leave the interface to the outside world as char *, because that
255     should make things easier for callers. We define a short type for unsigned char
256     to save lots of typing. I tried "uchar", but it causes problems on Digital
257     Unix, where it is defined in sys/types, so use "uschar" instead. */
258    
259     typedef unsigned char uschar;
260    
261     /* The real format of the start of the pcre block; the actual code vector
262     runs on as long as necessary after the end. */
263    
264     typedef struct real_pcre {
265     unsigned int magic_number;
266 nigel 25 const unsigned char *tables;
267 nigel 3 unsigned short int options;
268     unsigned char top_bracket;
269     unsigned char top_backref;
270     unsigned char first_char;
271     unsigned char code[1];
272     } real_pcre;
273    
274     /* The real format of the extra block returned by pcre_study(). */
275    
276     typedef struct real_pcre_extra {
277     unsigned char options;
278     unsigned char start_bits[32];
279     } real_pcre_extra;
280    
281    
282 nigel 25 /* Structure for passing "static" information around between the functions
283     doing the compiling, so that they are thread-safe. */
284 nigel 3
285 nigel 25 typedef struct compile_data {
286     const uschar *lcc; /* Points to lower casing table */
287     const uschar *fcc; /* Points to case-flippint table */
288     const uschar *cbits; /* Points to character type table */
289     const uschar *ctypes; /* Points to table of type maps */
290     } compile_data;
291 nigel 3
292 nigel 25 /* Structure for passing "static" information around between the functions
293     doing the matching, so that they are thread-safe. */
294    
295     typedef struct match_data {
296     int errorcode; /* As it says */
297     int *offset_vector; /* Offset vector */
298     int offset_end; /* One past the end */
299     int offset_max; /* The maximum usable for return data */
300     const uschar *lcc; /* Points to lower casing table */
301     const uschar *ctypes; /* Points to table of type maps */
302     BOOL offset_overflow; /* Set if too many extractions */
303     BOOL notbol; /* NOTBOL flag */
304     BOOL noteol; /* NOTEOL flag */
305     BOOL endonly; /* Dollar not before final \n */
306     const uschar *start_subject; /* Start of the subject string */
307     const uschar *end_subject; /* End of the subject string */
308     const uschar *end_match_ptr; /* Subject position at end match */
309     int end_offset_top; /* Highwater mark at end of match */
310     } match_data;
311    
312     /* Bit definitions for entries in the pcre_ctypes table. */
313    
314 nigel 3 #define ctype_space 0x01
315     #define ctype_letter 0x02
316     #define ctype_digit 0x04
317     #define ctype_xdigit 0x08
318     #define ctype_word 0x10 /* alphameric or '_' */
319     #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
320    
321 nigel 25 /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
322     of bits for a class map. */
323 nigel 3
324 nigel 25 #define cbit_digit 0 /* for \d */
325     #define cbit_word 32 /* for \w */
326     #define cbit_space 64 /* for \s */
327     #define cbit_length 96 /* Length of the cbits table */
328 nigel 3
329 nigel 25 /* Offsets of the various tables from the base tables pointer, and
330     total length. */
331    
332     #define lcc_offset 0
333     #define fcc_offset 256
334     #define cbits_offset 512
335     #define ctypes_offset (cbits_offset + cbit_length)
336     #define tables_length (ctypes_offset + 256)
337    
338 nigel 3 /* End of internal.h */

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12