1 /*************************************************
2 * Perl-Compatible Regular Expressions *
3 *************************************************/
4
5
6 /* 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
8 the file Tech.Notes for some information on the internals.
9
10 Written by: Philip Hazel <ph10@cam.ac.uk>
11
12 Copyright (c) 1997-2001 University of Cambridge
13
14 -----------------------------------------------------------------------------
15 Permission is granted to anyone to use this software for any purpose on any
16 computer system, and to redistribute it freely, subject to the following
17 restrictions:
18
19 1. This software is distributed in the hope that it will be useful,
20 but WITHOUT ANY WARRANTY; without even the implied warranty of
21 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
22
23 2. The origin of this software must not be misrepresented, either by
24 explicit claim or by omission.
25
26 3. Altered versions must be plainly marked as such, and must not be
27 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
36 modules, but which are not relevant to the outside. */
37
38 #include "pcre_config.h"
39
40 /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
41 define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
42 is set. Otherwise, include an emulating function for those systems that have
43 neither (there some non-Unix environments where this is the case). This assumes
44 that all calls to memmove are moving strings upwards in store, which is the
45 case in PCRE. */
46
47 #if ! HAVE_MEMMOVE
48 # undef memmove /* some systems may have a macro */
49 # if HAVE_BCOPY
50 # define memmove(a, b, c) bcopy(b, a, c)
51 # else
52 void *
pcre_memmove(unsigned char * dest,const unsigned char * src,size_t n)53 pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)
54 {
55 int i;
56 dest += n;
57 src += n;
58 for (i = 0; i < n; ++i) *(--dest) = *(--src);
59 }
60 # define memmove(a, b, c) pcre_memmove(a, b, c)
61 # endif
62 #endif
63
64 /* Standard C headers plus the external interface definition */
65
66 #include <ctype.h>
67 #include <limits.h>
68 #include <stddef.h>
69 #include <stdio.h>
70 #include <stdlib.h>
71 #include <string.h>
72 #include "pcre.h"
73
74 /* In case there is no definition of offsetof() provided - though any proper
75 Standard C system should have one. */
76
77 #ifndef offsetof
78 # define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
79 #endif
80
81 /* These are the public options that can change during matching. */
82
83 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
84
85 /* Private options flags start at the most significant end of the four bytes,
86 but skip the top bit so we can use ints for convenience without getting tangled
87 with negative values. The public options defined in pcre.h start at the least
88 significant end. Make sure they don't overlap, though now that we have expanded
89 to four bytes there is plenty of space. */
90
91 #define PCRE_FIRSTSET 0x40000000 /* first_char is set */
92 #define PCRE_REQCHSET 0x20000000 /* req_char is set */
93 #define PCRE_STARTLINE 0x10000000 /* start after \n for multiline */
94 #define PCRE_INGROUP 0x08000000 /* compiling inside a group */
95 #define PCRE_ICHANGED 0x04000000 /* i option changes within regex */
96
97 /* Options for the "extra" block produced by pcre_study(). */
98
99 #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
100
101 /* Masks for identifying the public options which are permitted at compile
102 time, run time or study time, respectively. */
103
104 #define PUBLIC_OPTIONS \
105 (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
106 PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8)
107
108 #define PUBLIC_EXEC_OPTIONS \
109 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY)
110
111 #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
112
113 /* Magic number to provide a small check against being handed junk. */
114
115 #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
116
117 /* Miscellaneous definitions */
118
119 typedef int BOOL;
120
121 #define FALSE 0
122 #define TRUE 1
123
124 /* Escape items that are just an encoding of a particular data value. Note that
125 ESC_N is defined as yet another macro, which is set in config.h to either \n
126 (the default) or \r (which some people want). */
127
128 #ifndef ESC_E
129 # define ESC_E 27
130 #endif
131
132 #ifndef ESC_F
133 # define ESC_F '\f'
134 #endif
135
136 #ifndef ESC_N
137 # define ESC_N NEWLINE
138 #endif
139
140 #ifndef ESC_R
141 # define ESC_R '\r'
142 #endif
143
144 #ifndef ESC_T
145 # define ESC_T '\t'
146 #endif
147
148 /* These are escaped items that aren't just an encoding of a particular data
149 value such as \n. They must have non-zero values, as check_escape() returns
150 their negation. Also, they must appear in the same order as in the opcode
151 definitions below, up to ESC_z. The final one must be ESC_REF as subsequent
152 values are used for \1, \2, \3, etc. There is a test in the code for an escape
153 greater than ESC_b and less than ESC_Z to detect the types that may be
154 repeated. If any new escapes are put in-between that don't consume a character,
155 that code will have to change. */
156
157 enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,
158 ESC_Z, ESC_z, ESC_REF };
159
160 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
161 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
162 OP_EOD must correspond in order to the list of escapes immediately above. */
163
164 enum {
165 OP_END, /* End of pattern */
166
167 /* Values corresponding to backslashed metacharacters */
168
169 OP_SOD, /* Start of data: \A */
170 OP_NOT_WORD_BOUNDARY, /* \B */
171 OP_WORD_BOUNDARY, /* \b */
172 OP_NOT_DIGIT, /* \D */
173 OP_DIGIT, /* \d */
174 OP_NOT_WHITESPACE, /* \S */
175 OP_WHITESPACE, /* \s */
176 OP_NOT_WORDCHAR, /* \W */
177 OP_WORDCHAR, /* \w */
178 OP_EODN, /* End of data or \n at end of data: \Z. */
179 OP_EOD, /* End of data: \z */
180
181 OP_OPT, /* Set runtime options */
182 OP_CIRC, /* Start of line - varies with multiline switch */
183 OP_DOLL, /* End of line - varies with multiline switch */
184 OP_ANY, /* Match any character */
185 OP_CHARS, /* Match string of characters */
186 OP_NOT, /* Match anything but the following char */
187
188 OP_STAR, /* The maximizing and minimizing versions of */
189 OP_MINSTAR, /* all these opcodes must come in pairs, with */
190 OP_PLUS, /* the minimizing one second. */
191 OP_MINPLUS, /* This first set applies to single characters */
192 OP_QUERY,
193 OP_MINQUERY,
194 OP_UPTO, /* From 0 to n matches */
195 OP_MINUPTO,
196 OP_EXACT, /* Exactly n matches */
197
198 OP_NOTSTAR, /* The maximizing and minimizing versions of */
199 OP_NOTMINSTAR, /* all these opcodes must come in pairs, with */
200 OP_NOTPLUS, /* the minimizing one second. */
201 OP_NOTMINPLUS, /* This first set applies to "not" single characters */
202 OP_NOTQUERY,
203 OP_NOTMINQUERY,
204 OP_NOTUPTO, /* From 0 to n matches */
205 OP_NOTMINUPTO,
206 OP_NOTEXACT, /* Exactly n matches */
207
208 OP_TYPESTAR, /* The maximizing and minimizing versions of */
209 OP_TYPEMINSTAR, /* all these opcodes must come in pairs, with */
210 OP_TYPEPLUS, /* the minimizing one second. These codes must */
211 OP_TYPEMINPLUS, /* be in exactly the same order as those above. */
212 OP_TYPEQUERY, /* This set applies to character types such as \d */
213 OP_TYPEMINQUERY,
214 OP_TYPEUPTO, /* From 0 to n matches */
215 OP_TYPEMINUPTO,
216 OP_TYPEEXACT, /* Exactly n matches */
217
218 OP_CRSTAR, /* The maximizing and minimizing versions of */
219 OP_CRMINSTAR, /* all these opcodes must come in pairs, with */
220 OP_CRPLUS, /* the minimizing one second. These codes must */
221 OP_CRMINPLUS, /* be in exactly the same order as those above. */
222 OP_CRQUERY, /* These are for character classes and back refs */
223 OP_CRMINQUERY,
224 OP_CRRANGE, /* These are different to the three seta above. */
225 OP_CRMINRANGE,
226
227 OP_CLASS, /* Match a character class */
228 OP_REF, /* Match a back reference */
229 OP_RECURSE, /* Match this pattern recursively */
230
231 OP_ALT, /* Start of alternation */
232 OP_KET, /* End of group that doesn't have an unbounded repeat */
233 OP_KETRMAX, /* These two must remain together and in this */
234 OP_KETRMIN, /* order. They are for groups the repeat for ever. */
235
236 /* The assertions must come before ONCE and COND */
237
238 OP_ASSERT, /* Positive lookahead */
239 OP_ASSERT_NOT, /* Negative lookahead */
240 OP_ASSERTBACK, /* Positive lookbehind */
241 OP_ASSERTBACK_NOT, /* Negative lookbehind */
242 OP_REVERSE, /* Move pointer back - used in lookbehind assertions */
243
244 /* ONCE and COND must come after the assertions, with ONCE first,
245 as there's a test for >= ONCE for a subpattern that isn't an assertion. */
246
247 OP_ONCE, /* Once matched, don't back up into the subpattern */
248 OP_COND, /* Conditional group */
249 OP_CREF, /* Used to hold an extraction string number (cond ref)*/
250
251 OP_BRAZERO, /* These two must remain together and in this */
252 OP_BRAMINZERO, /* order. */
253
254 OP_BRANUMBER, /* Used for extracting brackets whose number is greater
255 than can fit into an opcode. */
256
257 OP_BRA /* This and greater values are used for brackets that
258 extract substrings up to a basic limit. After that,
259 use is made of OP_BRANUMBER. */
260 };
261
262 /* The highest extraction number before we have to start using additional
263 bytes. (Originally PCRE didn't have support for extraction counts highter than
264 this number.) The value is limited by the number of opcodes left after OP_BRA,
265 i.e. 255 - OP_BRA. We actually set it a bit lower to leave room for additional
266 opcodes. */
267
268 #define EXTRACT_BASIC_MAX 150
269
270 /* The texts of compile-time error messages are defined as macros here so that
271 they can be accessed by the POSIX wrapper and converted into error codes. Yes,
272 I could have used error codes in the first place, but didn't feel like changing
273 just to accommodate the POSIX wrapper. */
274
275 #define ERR1 "\\ at end of pattern"
276 #define ERR2 "\\c at end of pattern"
277 #define ERR3 "unrecognized character follows \\"
278 #define ERR4 "numbers out of order in {} quantifier"
279 #define ERR5 "number too big in {} quantifier"
280 #define ERR6 "missing terminating ] for character class"
281 #define ERR7 "invalid escape sequence in character class"
282 #define ERR8 "range out of order in character class"
283 #define ERR9 "nothing to repeat"
284 #define ERR10 "operand of unlimited repeat could match the empty string"
285 #define ERR11 "internal error: unexpected repeat"
286 #define ERR12 "unrecognized character after (?"
287 #define ERR13 "unused error"
288 #define ERR14 "missing )"
289 #define ERR15 "back reference to non-existent subpattern"
290 #define ERR16 "erroffset passed as NULL"
291 #define ERR17 "unknown option bit(s) set"
292 #define ERR18 "missing ) after comment"
293 #define ERR19 "parentheses nested too deeply"
294 #define ERR20 "regular expression too large"
295 #define ERR21 "failed to get memory"
296 #define ERR22 "unmatched parentheses"
297 #define ERR23 "internal error: code overflow"
298 #define ERR24 "unrecognized character after (?<"
299 #define ERR25 "lookbehind assertion is not fixed length"
300 #define ERR26 "malformed number after (?("
301 #define ERR27 "conditional group contains more than two branches"
302 #define ERR28 "assertion expected after (?("
303 #define ERR29 "(?p must be followed by )"
304 #define ERR30 "unknown POSIX class name"
305 #define ERR31 "POSIX collating elements are not supported"
306 #define ERR32 "this version of PCRE is not compiled with PCRE_UTF8 support"
307 #define ERR33 "characters with values > 255 are not yet supported in classes"
308 #define ERR34 "character value in \\x{...} sequence is too large"
309 #define ERR35 "invalid condition (?(0)"
310
311 /* All character handling must be done as unsigned characters. Otherwise there
312 are problems with top-bit-set characters and functions such as isspace().
313 However, we leave the interface to the outside world as char *, because that
314 should make things easier for callers. We define a short type for unsigned char
315 to save lots of typing. I tried "uchar", but it causes problems on Digital
316 Unix, where it is defined in sys/types, so use "uschar" instead. */
317
318 typedef unsigned char uschar;
319
320 /* The real format of the start of the pcre block; the actual code vector
321 runs on as long as necessary after the end. */
322
323 typedef struct real_pcre {
324 unsigned long int magic_number;
325 size_t size;
326 const unsigned char *tables;
327 unsigned long int options;
328 unsigned short int top_bracket;
329 unsigned short int top_backref;
330 uschar first_char;
331 uschar req_char;
332 uschar code[1];
333 } real_pcre;
334
335 /* The real format of the extra block returned by pcre_study(). */
336
337 typedef struct real_pcre_extra {
338 uschar options;
339 uschar start_bits[32];
340 } real_pcre_extra;
341
342
343 /* Structure for passing "static" information around between the functions
344 doing the compiling, so that they are thread-safe. */
345
346 typedef struct compile_data {
347 const uschar *lcc; /* Points to lower casing table */
348 const uschar *fcc; /* Points to case-flipping table */
349 const uschar *cbits; /* Points to character type table */
350 const uschar *ctypes; /* Points to table of type maps */
351 } compile_data;
352
353 /* Structure for passing "static" information around between the functions
354 doing the matching, so that they are thread-safe. */
355
356 typedef struct match_data {
357 int errorcode; /* As it says */
358 int *offset_vector; /* Offset vector */
359 int offset_end; /* One past the end */
360 int offset_max; /* The maximum usable for return data */
361 const uschar *lcc; /* Points to lower casing table */
362 const uschar *ctypes; /* Points to table of type maps */
363 BOOL offset_overflow; /* Set if too many extractions */
364 BOOL notbol; /* NOTBOL flag */
365 BOOL noteol; /* NOTEOL flag */
366 BOOL utf8; /* UTF8 flag */
367 BOOL endonly; /* Dollar not before final \n */
368 BOOL notempty; /* Empty string match not wanted */
369 const uschar *start_pattern; /* For use when recursing */
370 const uschar *start_subject; /* Start of the subject string */
371 const uschar *end_subject; /* End of the subject string */
372 const uschar *start_match; /* Start of this match attempt */
373 const uschar *end_match_ptr; /* Subject position at end match */
374 int end_offset_top; /* Highwater mark at end of match */
375 } match_data;
376
377 /* Bit definitions for entries in the pcre_ctypes table. */
378
379 #define ctype_space 0x01
380 #define ctype_letter 0x02
381 #define ctype_digit 0x04
382 #define ctype_xdigit 0x08
383 #define ctype_word 0x10 /* alphameric or '_' */
384 #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
385
386 /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
387 of bits for a class map. Some classes are built by combining these tables. */
388
389 #define cbit_space 0 /* [:space:] or \s */
390 #define cbit_xdigit 32 /* [:xdigit:] */
391 #define cbit_digit 64 /* [:digit:] or \d */
392 #define cbit_upper 96 /* [:upper:] */
393 #define cbit_lower 128 /* [:lower:] */
394 #define cbit_word 160 /* [:word:] or \w */
395 #define cbit_graph 192 /* [:graph:] */
396 #define cbit_print 224 /* [:print:] */
397 #define cbit_punct 256 /* [:punct:] */
398 #define cbit_cntrl 288 /* [:cntrl:] */
399 #define cbit_length 320 /* Length of the cbits table */
400
401 /* Offsets of the various tables from the base tables pointer, and
402 total length. */
403
404 #define lcc_offset 0
405 #define fcc_offset 256
406 #define cbits_offset 512
407 #define ctypes_offset (cbits_offset + cbit_length)
408 #define tables_length (ctypes_offset + 256)
409
410 /* End of internal.h */
411