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

Contents of /code/trunk/internal.h

Parent Directory Parent Directory | Revision Log Revision Log


Revision 37 - (show annotations) (download)
Sat Feb 24 21:39:09 2007 UTC (7 years, 9 months ago) by nigel
File MIME type: text/plain
File size: 13979 byte(s)
Load pcre-2.07 into code/trunk.

1 /*************************************************
2 * Perl-Compatible Regular Expressions *
3 *************************************************/
4
5
6 #define PCRE_VERSION "2.07 29-Jul-1999"
7
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 Copyright (c) 1997-1999 University of Cambridge
16
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
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 -----------------------------------------------------------------------------
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 #undef memmove /* some systems may have a macro */
46 #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 /* In case there is no definition of offsetof() provided - though any proper
60 Standard C system should have one. */
61
62 #ifndef offsetof
63 #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
64 #endif
65
66 /* These are the public options that can change during matching. */
67
68 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
69
70 /* Private options flags start at the most significant end of the four bytes,
71 but skip the top bit so we can use ints for convenience without getting tangled
72 with negative values. The public options defined in pcre.h start at the least
73 significant end. Make sure they don't overlap, though now that we have expanded
74 to four bytes there is plenty of space. */
75
76 #define PCRE_FIRSTSET 0x40000000 /* first_char is set */
77 #define PCRE_REQCHSET 0x20000000 /* req_char is set */
78 #define PCRE_STARTLINE 0x10000000 /* start after \n for multiline */
79 #define PCRE_INGROUP 0x08000000 /* compiling inside a group */
80 #define PCRE_ICHANGED 0x04000000 /* i option changes within regex */
81
82 /* Options for the "extra" block produced by pcre_study(). */
83
84 #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
85
86 /* Masks for identifying the public options which are permitted at compile
87 time, run time or study time, respectively. */
88
89 #define PUBLIC_OPTIONS \
90 (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
91 PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY)
92
93 #define PUBLIC_EXEC_OPTIONS \
94 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY)
95
96 #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
97
98 /* Magic number to provide a small check against being handed junk. */
99
100 #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
101
102 /* Miscellaneous definitions */
103
104 typedef int BOOL;
105
106 #define FALSE 0
107 #define TRUE 1
108
109 /* These are escaped items that aren't just an encoding of a particular data
110 value such as \n. They must have non-zero values, as check_escape() returns
111 their negation. Also, they must appear in the same order as in the opcode
112 definitions below, up to ESC_z. The final one must be ESC_REF as subsequent
113 values are used for \1, \2, \3, etc. There is a test in the code for an escape
114 greater than ESC_b and less than ESC_X to detect the types that may be
115 repeated. If any new escapes are put in-between that don't consume a character,
116 that code will have to change. */
117
118 enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,
119 ESC_Z, ESC_z, ESC_REF };
120
121 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
122 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
123 OP_EOD must correspond in order to the list of escapes immediately above. */
124
125 enum {
126 OP_END, /* End of pattern */
127
128 /* Values corresponding to backslashed metacharacters */
129
130 OP_SOD, /* Start of data: \A */
131 OP_NOT_WORD_BOUNDARY, /* \B */
132 OP_WORD_BOUNDARY, /* \b */
133 OP_NOT_DIGIT, /* \D */
134 OP_DIGIT, /* \d */
135 OP_NOT_WHITESPACE, /* \S */
136 OP_WHITESPACE, /* \s */
137 OP_NOT_WORDCHAR, /* \W */
138 OP_WORDCHAR, /* \w */
139 OP_EODN, /* End of data or \n at end of data: \Z. */
140 OP_EOD, /* End of data: \z */
141
142 OP_OPT, /* Set runtime options */
143 OP_CIRC, /* Start of line - varies with multiline switch */
144 OP_DOLL, /* End of line - varies with multiline switch */
145 OP_ANY, /* Match any character */
146 OP_CHARS, /* Match string of characters */
147 OP_NOT, /* Match anything but the following char */
148
149 OP_STAR, /* The maximizing and minimizing versions of */
150 OP_MINSTAR, /* all these opcodes must come in pairs, with */
151 OP_PLUS, /* the minimizing one second. */
152 OP_MINPLUS, /* This first set applies to single characters */
153 OP_QUERY,
154 OP_MINQUERY,
155 OP_UPTO, /* From 0 to n matches */
156 OP_MINUPTO,
157 OP_EXACT, /* Exactly n matches */
158
159 OP_NOTSTAR, /* The maximizing and minimizing versions of */
160 OP_NOTMINSTAR, /* all these opcodes must come in pairs, with */
161 OP_NOTPLUS, /* the minimizing one second. */
162 OP_NOTMINPLUS, /* This first set applies to "not" single characters */
163 OP_NOTQUERY,
164 OP_NOTMINQUERY,
165 OP_NOTUPTO, /* From 0 to n matches */
166 OP_NOTMINUPTO,
167 OP_NOTEXACT, /* Exactly n matches */
168
169 OP_TYPESTAR, /* The maximizing and minimizing versions of */
170 OP_TYPEMINSTAR, /* all these opcodes must come in pairs, with */
171 OP_TYPEPLUS, /* the minimizing one second. These codes must */
172 OP_TYPEMINPLUS, /* be in exactly the same order as those above. */
173 OP_TYPEQUERY, /* This set applies to character types such as \d */
174 OP_TYPEMINQUERY,
175 OP_TYPEUPTO, /* From 0 to n matches */
176 OP_TYPEMINUPTO,
177 OP_TYPEEXACT, /* Exactly n matches */
178
179 OP_CRSTAR, /* The maximizing and minimizing versions of */
180 OP_CRMINSTAR, /* all these opcodes must come in pairs, with */
181 OP_CRPLUS, /* the minimizing one second. These codes must */
182 OP_CRMINPLUS, /* be in exactly the same order as those above. */
183 OP_CRQUERY, /* These are for character classes and back refs */
184 OP_CRMINQUERY,
185 OP_CRRANGE, /* These are different to the three seta above. */
186 OP_CRMINRANGE,
187
188 OP_CLASS, /* Match a character class */
189 OP_REF, /* Match a back reference */
190
191 OP_ALT, /* Start of alternation */
192 OP_KET, /* End of group that doesn't have an unbounded repeat */
193 OP_KETRMAX, /* These two must remain together and in this */
194 OP_KETRMIN, /* order. They are for groups the repeat for ever. */
195
196 /* The assertions must come before ONCE and COND */
197
198 OP_ASSERT, /* Positive lookahead */
199 OP_ASSERT_NOT, /* Negative lookahead */
200 OP_ASSERTBACK, /* Positive lookbehind */
201 OP_ASSERTBACK_NOT, /* Negative lookbehind */
202 OP_REVERSE, /* Move pointer back - used in lookbehind assertions */
203
204 /* ONCE and COND must come after the assertions, with ONCE first, as there's
205 a test for >= ONCE for a subpattern that isn't an assertion. */
206
207 OP_ONCE, /* Once matched, don't back up into the subpattern */
208 OP_COND, /* Conditional group */
209 OP_CREF, /* Used to hold an extraction string number */
210
211 OP_BRAZERO, /* These two must remain together and in this */
212 OP_BRAMINZERO, /* order. */
213
214 OP_BRA /* This and greater values are used for brackets that
215 extract substrings. */
216 };
217
218 /* The highest extraction number. This is limited by the number of opcodes
219 left after OP_BRA, i.e. 255 - OP_BRA. We actually set it somewhat lower. */
220
221 #define EXTRACT_MAX 99
222
223 /* The texts of compile-time error messages are defined as macros here so that
224 they can be accessed by the POSIX wrapper and converted into error codes. Yes,
225 I could have used error codes in the first place, but didn't feel like changing
226 just to accommodate the POSIX wrapper. */
227
228 #define ERR1 "\\ at end of pattern"
229 #define ERR2 "\\c at end of pattern"
230 #define ERR3 "unrecognized character follows \\"
231 #define ERR4 "numbers out of order in {} quantifier"
232 #define ERR5 "number too big in {} quantifier"
233 #define ERR6 "missing terminating ] for character class"
234 #define ERR7 "invalid escape sequence in character class"
235 #define ERR8 "range out of order in character class"
236 #define ERR9 "nothing to repeat"
237 #define ERR10 "operand of unlimited repeat could match the empty string"
238 #define ERR11 "internal error: unexpected repeat"
239 #define ERR12 "unrecognized character after (?"
240 #define ERR13 "too many capturing parenthesized sub-patterns"
241 #define ERR14 "missing )"
242 #define ERR15 "back reference to non-existent subpattern"
243 #define ERR16 "erroffset passed as NULL"
244 #define ERR17 "unknown option bit(s) set"
245 #define ERR18 "missing ) after comment"
246 #define ERR19 "too many sets of parentheses"
247 #define ERR20 "regular expression too large"
248 #define ERR21 "failed to get memory"
249 #define ERR22 "unmatched parentheses"
250 #define ERR23 "internal error: code overflow"
251 #define ERR24 "unrecognized character after (?<"
252 #define ERR25 "lookbehind assertion is not fixed length"
253 #define ERR26 "malformed number after (?("
254 #define ERR27 "conditional group contains more than two branches"
255 #define ERR28 "assertion expected after (?("
256
257 /* All character handling must be done as unsigned characters. Otherwise there
258 are problems with top-bit-set characters and functions such as isspace().
259 However, we leave the interface to the outside world as char *, because that
260 should make things easier for callers. We define a short type for unsigned char
261 to save lots of typing. I tried "uchar", but it causes problems on Digital
262 Unix, where it is defined in sys/types, so use "uschar" instead. */
263
264 typedef unsigned char uschar;
265
266 /* The real format of the start of the pcre block; the actual code vector
267 runs on as long as necessary after the end. */
268
269 typedef struct real_pcre {
270 unsigned long int magic_number;
271 const unsigned char *tables;
272 unsigned long int options;
273 uschar top_bracket;
274 uschar top_backref;
275 uschar first_char;
276 uschar req_char;
277 uschar code[1];
278 } real_pcre;
279
280 /* The real format of the extra block returned by pcre_study(). */
281
282 typedef struct real_pcre_extra {
283 uschar options;
284 uschar start_bits[32];
285 } real_pcre_extra;
286
287
288 /* Structure for passing "static" information around between the functions
289 doing the compiling, so that they are thread-safe. */
290
291 typedef struct compile_data {
292 const uschar *lcc; /* Points to lower casing table */
293 const uschar *fcc; /* Points to case-flipping table */
294 const uschar *cbits; /* Points to character type table */
295 const uschar *ctypes; /* Points to table of type maps */
296 } compile_data;
297
298 /* Structure for passing "static" information around between the functions
299 doing the matching, so that they are thread-safe. */
300
301 typedef struct match_data {
302 int errorcode; /* As it says */
303 int *offset_vector; /* Offset vector */
304 int offset_end; /* One past the end */
305 int offset_max; /* The maximum usable for return data */
306 const uschar *lcc; /* Points to lower casing table */
307 const uschar *ctypes; /* Points to table of type maps */
308 BOOL offset_overflow; /* Set if too many extractions */
309 BOOL notbol; /* NOTBOL flag */
310 BOOL noteol; /* NOTEOL flag */
311 BOOL endonly; /* Dollar not before final \n */
312 BOOL notempty; /* Empty string match not wanted */
313 const uschar *start_subject; /* Start of the subject string */
314 const uschar *end_subject; /* End of the subject string */
315 const uschar *start_match; /* Start of this match attempt */
316 const uschar *end_match_ptr; /* Subject position at end match */
317 int end_offset_top; /* Highwater mark at end of match */
318 } match_data;
319
320 /* Bit definitions for entries in the pcre_ctypes table. */
321
322 #define ctype_space 0x01
323 #define ctype_letter 0x02
324 #define ctype_digit 0x04
325 #define ctype_xdigit 0x08
326 #define ctype_word 0x10 /* alphameric or '_' */
327 #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
328
329 /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
330 of bits for a class map. */
331
332 #define cbit_digit 0 /* for \d */
333 #define cbit_word 32 /* for \w */
334 #define cbit_space 64 /* for \s */
335 #define cbit_length 96 /* Length of the cbits table */
336
337 /* Offsets of the various tables from the base tables pointer, and
338 total length. */
339
340 #define lcc_offset 0
341 #define fcc_offset 256
342 #define cbits_offset 512
343 #define ctypes_offset (cbits_offset + cbit_length)
344 #define tables_length (ctypes_offset + 256)
345
346 /* End of internal.h */

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12