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) 1997-2000 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 /* 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(),
43 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 #if ! HAVE_MEMMOVE
50 #undef memmove /* some systems may have a macro */
51 #if HAVE_BCOPY
52 #define memmove(a, b, c) bcopy(b, a, c)
53 #else
54 void *
pcre_memmove(unsigned char * dest,const unsigned char * src,size_t n)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
65
66 /* Standard C headers plus the external interface definition */
67
68 #include <ctype.h>
69 #include <limits.h>
70 #include <stddef.h>
71 #include <stdio.h>
72 #include <stdlib.h>
73 #include <string.h>
74 #include "pcre.h"
75
76 /* In case there is no definition of offsetof() provided - though any proper
77 Standard C system should have one. */
78
79 #ifndef offsetof
80 #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
81 #endif
82
83 /* These are the public options that can change during matching. */
84
85 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
86
87 /* Private options flags start at the most significant end of the four bytes,
88 but skip the top bit so we can use ints for convenience without getting tangled
89 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 to four bytes there is plenty of space. */
92
93 #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(). */
100
101 #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
102
103 /* Masks for identifying the public options which are permitted at compile
104 time, run time or study time, respectively. */
105
106 #define PUBLIC_OPTIONS \
107 (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
108 PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8)
109
110 #define PUBLIC_EXEC_OPTIONS \
111 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY)
112
113 #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
114
115 /* Magic number to provide a small check against being handed junk. */
116
117 #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
118
119 /* Miscellaneous definitions */
120
121 typedef int BOOL;
122
123 #define FALSE 0
124 #define TRUE 1
125
126 /* These are escaped items that aren't just an encoding of a particular data
127 value such as \n. They must have non-zero values, as check_escape() returns
128 their negation. Also, they must appear in the same order as in the opcode
129 definitions below, up to ESC_z. The final one must be ESC_REF as subsequent
130 values are used for \1, \2, \3, etc. There is a test in the code for an escape
131 greater than ESC_b and less than ESC_X to detect the types that may be
132 repeated. If any new escapes are put in-between that don't consume a character,
133 that code will have to change. */
134
135 enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,
136 ESC_Z, ESC_z, ESC_REF };
137
138 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
139 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
140 OP_EOD must correspond in order to the list of escapes immediately above. */
141
142 enum {
143 OP_END, /* End of pattern */
144
145 /* Values corresponding to backslashed metacharacters */
146
147 OP_SOD, /* Start of data: \A */
148 OP_NOT_WORD_BOUNDARY, /* \B */
149 OP_WORD_BOUNDARY, /* \b */
150 OP_NOT_DIGIT, /* \D */
151 OP_DIGIT, /* \d */
152 OP_NOT_WHITESPACE, /* \S */
153 OP_WHITESPACE, /* \s */
154 OP_NOT_WORDCHAR, /* \W */
155 OP_WORDCHAR, /* \w */
156 OP_EODN, /* End of data or \n at end of data: \Z. */
157 OP_EOD, /* End of data: \z */
158
159 OP_OPT, /* Set runtime options */
160 OP_CIRC, /* Start of line - varies with multiline switch */
161 OP_DOLL, /* End of line - varies with multiline switch */
162 OP_ANY, /* Match any character */
163 OP_CHARS, /* Match string of characters */
164 OP_NOT, /* Match anything but the following char */
165
166 OP_STAR, /* The maximizing and minimizing versions of */
167 OP_MINSTAR, /* all these opcodes must come in pairs, with */
168 OP_PLUS, /* the minimizing one second. */
169 OP_MINPLUS, /* This first set applies to single characters */
170 OP_QUERY,
171 OP_MINQUERY,
172 OP_UPTO, /* From 0 to n matches */
173 OP_MINUPTO,
174 OP_EXACT, /* Exactly n matches */
175
176 OP_NOTSTAR, /* The maximizing and minimizing versions of */
177 OP_NOTMINSTAR, /* all these opcodes must come in pairs, with */
178 OP_NOTPLUS, /* the minimizing one second. */
179 OP_NOTMINPLUS, /* This first set applies to "not" single characters */
180 OP_NOTQUERY,
181 OP_NOTMINQUERY,
182 OP_NOTUPTO, /* From 0 to n matches */
183 OP_NOTMINUPTO,
184 OP_NOTEXACT, /* Exactly n matches */
185
186 OP_TYPESTAR, /* The maximizing and minimizing versions of */
187 OP_TYPEMINSTAR, /* all these opcodes must come in pairs, with */
188 OP_TYPEPLUS, /* the minimizing one second. These codes must */
189 OP_TYPEMINPLUS, /* be in exactly the same order as those above. */
190 OP_TYPEQUERY, /* This set applies to character types such as \d */
191 OP_TYPEMINQUERY,
192 OP_TYPEUPTO, /* From 0 to n matches */
193 OP_TYPEMINUPTO,
194 OP_TYPEEXACT, /* Exactly n matches */
195
196 OP_CRSTAR, /* The maximizing and minimizing versions of */
197 OP_CRMINSTAR, /* all these opcodes must come in pairs, with */
198 OP_CRPLUS, /* the minimizing one second. These codes must */
199 OP_CRMINPLUS, /* be in exactly the same order as those above. */
200 OP_CRQUERY, /* These are for character classes and back refs */
201 OP_CRMINQUERY,
202 OP_CRRANGE, /* These are different to the three seta above. */
203 OP_CRMINRANGE,
204
205 OP_CLASS, /* Match a character class */
206 OP_REF, /* Match a back reference */
207 OP_RECURSE, /* Match this pattern recursively */
208
209 OP_ALT, /* Start of alternation */
210 OP_KET, /* End of group that doesn't have an unbounded repeat */
211 OP_KETRMAX, /* These two must remain together and in this */
212 OP_KETRMIN, /* order. They are for groups the repeat for ever. */
213
214 /* The assertions must come before ONCE and COND */
215
216 OP_ASSERT, /* Positive lookahead */
217 OP_ASSERT_NOT, /* Negative lookahead */
218 OP_ASSERTBACK, /* Positive lookbehind */
219 OP_ASSERTBACK_NOT, /* Negative lookbehind */
220 OP_REVERSE, /* Move pointer back - used in lookbehind assertions */
221
222 /* ONCE and COND must come after the assertions, with ONCE first, as there's
223 a test for >= ONCE for a subpattern that isn't an assertion. */
224
225 OP_ONCE, /* Once matched, don't back up into the subpattern */
226 OP_COND, /* Conditional group */
227 OP_CREF, /* Used to hold an extraction string number */
228
229 OP_BRAZERO, /* These two must remain together and in this */
230 OP_BRAMINZERO, /* order. */
231
232 OP_BRA /* This and greater values are used for brackets that
233 extract substrings. */
234 };
235
236 /* The highest extraction number. This is limited by the number of opcodes
237 left after OP_BRA, i.e. 255 - OP_BRA. We actually set it somewhat lower. */
238
239 #define EXTRACT_MAX 99
240
241 /* The texts of compile-time error messages are defined as macros here so that
242 they can be accessed by the POSIX wrapper and converted into error codes. Yes,
243 I could have used error codes in the first place, but didn't feel like changing
244 just to accommodate the POSIX wrapper. */
245
246 #define ERR1 "\\ at end of pattern"
247 #define ERR2 "\\c at end of pattern"
248 #define ERR3 "unrecognized character follows \\"
249 #define ERR4 "numbers out of order in {} quantifier"
250 #define ERR5 "number too big in {} quantifier"
251 #define ERR6 "missing terminating ] for character class"
252 #define ERR7 "invalid escape sequence in character class"
253 #define ERR8 "range out of order in character class"
254 #define ERR9 "nothing to repeat"
255 #define ERR10 "operand of unlimited repeat could match the empty string"
256 #define ERR11 "internal error: unexpected repeat"
257 #define ERR12 "unrecognized character after (?"
258 #define ERR13 "too many capturing parenthesized sub-patterns"
259 #define ERR14 "missing )"
260 #define ERR15 "back reference to non-existent subpattern"
261 #define ERR16 "erroffset passed as NULL"
262 #define ERR17 "unknown option bit(s) set"
263 #define ERR18 "missing ) after comment"
264 #define ERR19 "too many sets of parentheses"
265 #define ERR20 "regular expression too large"
266 #define ERR21 "failed to get memory"
267 #define ERR22 "unmatched parentheses"
268 #define ERR23 "internal error: code overflow"
269 #define ERR24 "unrecognized character after (?<"
270 #define ERR25 "lookbehind assertion is not fixed length"
271 #define ERR26 "malformed number after (?("
272 #define ERR27 "conditional group contains more than two branches"
273 #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 #define ERR35 "invalid condition (?(0)"
281
282 /* All character handling must be done as unsigned characters. Otherwise there
283 are problems with top-bit-set characters and functions such as isspace().
284 However, we leave the interface to the outside world as char *, because that
285 should make things easier for callers. We define a short type for unsigned char
286 to save lots of typing. I tried "uchar", but it causes problems on Digital
287 Unix, where it is defined in sys/types, so use "uschar" instead. */
288
289 typedef unsigned char uschar;
290
291 /* The real format of the start of the pcre block; the actual code vector
292 runs on as long as necessary after the end. */
293
294 typedef struct real_pcre {
295 unsigned long int magic_number;
296 size_t size;
297 const unsigned char *tables;
298 unsigned long int options;
299 uschar top_bracket;
300 uschar top_backref;
301 uschar first_char;
302 uschar req_char;
303 uschar code[1];
304 } real_pcre;
305
306 /* The real format of the extra block returned by pcre_study(). */
307
308 typedef struct real_pcre_extra {
309 uschar options;
310 uschar start_bits[32];
311 } real_pcre_extra;
312
313
314 /* Structure for passing "static" information around between the functions
315 doing the compiling, so that they are thread-safe. */
316
317 typedef struct compile_data {
318 const uschar *lcc; /* Points to lower casing table */
319 const uschar *fcc; /* Points to case-flipping table */
320 const uschar *cbits; /* Points to character type table */
321 const uschar *ctypes; /* Points to table of type maps */
322 } compile_data;
323
324 /* Structure for passing "static" information around between the functions
325 doing the matching, so that they are thread-safe. */
326
327 typedef struct match_data {
328 int errorcode; /* As it says */
329 int *offset_vector; /* Offset vector */
330 int offset_end; /* One past the end */
331 int offset_max; /* The maximum usable for return data */
332 const uschar *lcc; /* Points to lower casing table */
333 const uschar *ctypes; /* Points to table of type maps */
334 BOOL offset_overflow; /* Set if too many extractions */
335 BOOL notbol; /* NOTBOL flag */
336 BOOL noteol; /* NOTEOL flag */
337 BOOL utf8; /* UTF8 flag */
338 BOOL endonly; /* Dollar not before final \n */
339 BOOL notempty; /* Empty string match not wanted */
340 const uschar *start_pattern; /* For use when recursing */
341 const uschar *start_subject; /* Start of the subject string */
342 const uschar *end_subject; /* End of the subject string */
343 const uschar *start_match; /* Start of this match attempt */
344 const uschar *end_match_ptr; /* Subject position at end match */
345 int end_offset_top; /* Highwater mark at end of match */
346 } match_data;
347
348 /* Bit definitions for entries in the pcre_ctypes table. */
349
350 #define ctype_space 0x01
351 #define ctype_letter 0x02
352 #define ctype_digit 0x04
353 #define ctype_xdigit 0x08
354 #define ctype_word 0x10 /* alphameric or '_' */
355 #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
356
357 /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
358 of bits for a class map. Some classes are built by combining these tables. */
359
360 #define cbit_space 0 /* [:space:] or \s */
361 #define cbit_xdigit 32 /* [:xdigit:] */
362 #define cbit_digit 64 /* [:digit:] or \d */
363 #define cbit_upper 96 /* [:upper:] */
364 #define cbit_lower 128 /* [:lower:] */
365 #define cbit_word 160 /* [:word:] or \w */
366 #define cbit_graph 192 /* [:graph:] */
367 #define cbit_print 224 /* [:print:] */
368 #define cbit_punct 256 /* [:punct:] */
369 #define cbit_cntrl 288 /* [:cntrl:] */
370 #define cbit_length 320 /* Length of the cbits table */
371
372 /* Offsets of the various tables from the base tables pointer, and
373 total length. */
374
375 #define lcc_offset 0
376 #define fcc_offset 256
377 #define cbits_offset 512
378 #define ctypes_offset (cbits_offset + cbit_length)
379 #define tables_length (ctypes_offset + 256)
380
381 /* End of internal.h */
382