1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4 Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
5
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2, or (at your option)
9 any later version.
10
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
15
16 You should have received a copy of the GNU General Public License along
17 with this program; if not, write to the Free Software Foundation,
18 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
19
20 #ifndef _REGEX_INTERNAL_H
21 #define _REGEX_INTERNAL_H 1
22
23 #include <assert.h>
24 #include <ctype.h>
25 #include <stdbool.h>
26 #include <stdio.h>
27 #include <stdlib.h>
28 #include <string.h>
29
30 #if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET || defined _LIBC
31 # include <langinfo.h>
32 #endif
33 #if defined HAVE_LOCALE_H || defined _LIBC
34 # include <locale.h>
35 #endif
36 #if defined HAVE_WCHAR_H || defined _LIBC
37 # include <wchar.h>
38 #endif /* HAVE_WCHAR_H || _LIBC */
39 #if defined HAVE_WCTYPE_H || defined _LIBC
40 # include <wctype.h>
41 #endif /* HAVE_WCTYPE_H || _LIBC */
42 #if defined _LIBC
43 # include <bits/libc-lock.h>
44 #else
45 # define __libc_lock_define(CLASS,NAME)
46 # define __libc_lock_init(NAME) do { } while (0)
47 # define __libc_lock_lock(NAME) do { } while (0)
48 # define __libc_lock_unlock(NAME) do { } while (0)
49 #endif
50
51 /* In case that the system doesn't have isblank(). */
52 #if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank
53 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
54 #endif
55
56 #ifdef _LIBC
57 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
58 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
59 # include <locale/localeinfo.h>
60 # include <locale/elem-hash.h>
61 # include <locale/coll-lookup.h>
62 # endif
63 #endif
64
65 /* This is for other GNU distributions with internationalized messages. */
66 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
67 # include <libintl.h>
68 # ifdef _LIBC
69 # undef gettext
70 # define gettext(msgid) \
71 INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES)
72 # endif
73 #else
74 # define gettext(msgid) (msgid)
75 #endif
76
77 #ifndef gettext_noop
78 /* This define is so xgettext can find the internationalizable
79 strings. */
80 # define gettext_noop(String) String
81 #endif
82
83 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC
84 # define RE_ENABLE_I18N
85 #endif
86
87 #if __GNUC__ >= 3
88 # define BE(expr, val) __builtin_expect (expr, val)
89 #else
90 # define BE(expr, val) (expr)
91 #endif
92
93 /* Number of single byte character. */
94 #define SBC_MAX 256
95
96 #define COLL_ELEM_LEN_MAX 8
97
98 /* The character which represents newline. */
99 #define NEWLINE_CHAR '\n'
100 #define WIDE_NEWLINE_CHAR L'\n'
101
102 /* Rename to standard API for using out of glibc. */
103 #ifndef _LIBC
104 # define __wctype wctype
105 # define __iswctype iswctype
106 # define __btowc btowc
107 # ifndef __mempcpy
108 # define __mempcpy mempcpy
109 # endif
110 # define __wcrtomb wcrtomb
111 # define __regfree regfree
112 # define attribute_hidden
113 #endif /* not _LIBC */
114
115 #if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
116 # define __attribute(arg) __attribute__ (arg)
117 #else
118 # define __attribute(arg)
119 #endif
120
121 extern const char __re_error_msgid[] attribute_hidden;
122 extern const size_t __re_error_msgid_idx[] attribute_hidden;
123
124 typedef __re_idx_t Idx;
125
126 /* Special return value for failure to match. */
127 #define REG_MISSING ((Idx) -1)
128
129 /* Special return value for internal error. */
130 #define REG_ERROR ((Idx) -2)
131
132 /* Test whether N is a valid index, and is not one of the above. */
133 #ifdef _REGEX_LARGE_OFFSETS
134 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
135 #else
136 # define REG_VALID_INDEX(n) (0 <= (n))
137 #endif
138
139 /* Test whether N is a valid nonzero index. */
140 #ifdef _REGEX_LARGE_OFFSETS
141 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
142 #else
143 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
144 #endif
145
146 /* A hash value, suitable for computing hash tables. */
147 typedef __re_size_t re_hashval_t;
148
149 /* An integer used to represent a set of bits. It must be unsigned,
150 and must be at least as wide as unsigned int. */
151 typedef unsigned long int bitset_word;
152
153 /* Maximum value of a bitset word. It must be useful in preprocessor
154 contexts, and must be consistent with bitset_word. */
155 #define BITSET_WORD_MAX ULONG_MAX
156
157 /* Number of bits in a bitset word. Avoid greater-than-32-bit
158 integers and unconditional shifts by more than 31 bits, as they're
159 not portable. */
160 #if BITSET_WORD_MAX == 0xffffffff
161 # define BITSET_WORD_BITS 32
162 #elif BITSET_WORD_MAX >> 31 >> 5 == 1
163 # define BITSET_WORD_BITS 36
164 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
165 # define BITSET_WORD_BITS 48
166 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
167 # define BITSET_WORD_BITS 60
168 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
169 # define BITSET_WORD_BITS 64
170 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
171 # define BITSET_WORD_BITS 72
172 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
173 # define BITSET_WORD_BITS 128
174 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
175 # define BITSET_WORD_BITS 256
176 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
177 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
178 # if BITSET_WORD_BITS <= SBC_MAX
179 # error "Invalid SBC_MAX"
180 # endif
181 #else
182 # error "Add case for new bitset_word size"
183 #endif
184
185 /* Number of bitset words in a bitset. */
186 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
187
188 typedef bitset_word bitset[BITSET_WORDS];
189 typedef bitset_word *re_bitset_ptr_t;
190 typedef const bitset_word *re_const_bitset_ptr_t;
191
192 #define PREV_WORD_CONSTRAINT 0x0001
193 #define PREV_NOTWORD_CONSTRAINT 0x0002
194 #define NEXT_WORD_CONSTRAINT 0x0004
195 #define NEXT_NOTWORD_CONSTRAINT 0x0008
196 #define PREV_NEWLINE_CONSTRAINT 0x0010
197 #define NEXT_NEWLINE_CONSTRAINT 0x0020
198 #define PREV_BEGBUF_CONSTRAINT 0x0040
199 #define NEXT_ENDBUF_CONSTRAINT 0x0080
200 #define WORD_DELIM_CONSTRAINT 0x0100
201 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
202
203 typedef enum
204 {
205 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
206 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
207 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
208 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
209 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
210 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
211 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
212 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
213 WORD_DELIM = WORD_DELIM_CONSTRAINT,
214 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
215 } re_context_type;
216
217 typedef struct
218 {
219 Idx alloc;
220 Idx nelem;
221 Idx *elems;
222 } re_node_set;
223
224 typedef enum
225 {
226 NON_TYPE = 0,
227
228 /* Node type, These are used by token, node, tree. */
229 CHARACTER = 1,
230 END_OF_RE = 2,
231 SIMPLE_BRACKET = 3,
232 OP_BACK_REF = 4,
233 OP_PERIOD = 5,
234 #ifdef RE_ENABLE_I18N
235 COMPLEX_BRACKET = 6,
236 OP_UTF8_PERIOD = 7,
237 #endif /* RE_ENABLE_I18N */
238
239 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
240 when the debugger shows values of this enum type. */
241 #define EPSILON_BIT 8
242 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
243 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
244 OP_ALT = EPSILON_BIT | 2,
245 OP_DUP_ASTERISK = EPSILON_BIT | 3,
246 ANCHOR = EPSILON_BIT | 4,
247
248 /* Tree type, these are used only by tree. */
249 CONCAT = 16,
250 SUBEXP = 17,
251
252 /* Token type, these are used only by token. */
253 OP_DUP_PLUS = 18,
254 OP_DUP_QUESTION,
255 OP_OPEN_BRACKET,
256 OP_CLOSE_BRACKET,
257 OP_CHARSET_RANGE,
258 OP_OPEN_DUP_NUM,
259 OP_CLOSE_DUP_NUM,
260 OP_NON_MATCH_LIST,
261 OP_OPEN_COLL_ELEM,
262 OP_CLOSE_COLL_ELEM,
263 OP_OPEN_EQUIV_CLASS,
264 OP_CLOSE_EQUIV_CLASS,
265 OP_OPEN_CHAR_CLASS,
266 OP_CLOSE_CHAR_CLASS,
267 OP_WORD,
268 OP_NOTWORD,
269 OP_SPACE,
270 OP_NOTSPACE,
271 BACK_SLASH
272
273 } re_token_type_t;
274
275 #ifdef RE_ENABLE_I18N
276 typedef struct
277 {
278 /* Multibyte characters. */
279 wchar_t *mbchars;
280
281 /* Collating symbols. */
282 # ifdef _LIBC
283 int32_t *coll_syms;
284 # endif
285
286 /* Equivalence classes. */
287 # ifdef _LIBC
288 int32_t *equiv_classes;
289 # endif
290
291 /* Range expressions. */
292 # ifdef _LIBC
293 uint32_t *range_starts;
294 uint32_t *range_ends;
295 # else /* not _LIBC */
296 wchar_t *range_starts;
297 wchar_t *range_ends;
298 # endif /* not _LIBC */
299
300 /* Character classes. */
301 wctype_t *char_classes;
302
303 /* If this character set is the non-matching list. */
304 unsigned int non_match : 1;
305
306 /* # of multibyte characters. */
307 Idx nmbchars;
308
309 /* # of collating symbols. */
310 Idx ncoll_syms;
311
312 /* # of equivalence classes. */
313 Idx nequiv_classes;
314
315 /* # of range expressions. */
316 Idx nranges;
317
318 /* # of character classes. */
319 Idx nchar_classes;
320 } re_charset_t;
321 #endif /* RE_ENABLE_I18N */
322
323 typedef struct
324 {
325 union
326 {
327 unsigned char c; /* for CHARACTER */
328 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
329 #ifdef RE_ENABLE_I18N
330 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
331 #endif /* RE_ENABLE_I18N */
332 Idx idx; /* for BACK_REF */
333 re_context_type ctx_type; /* for ANCHOR */
334 } opr;
335 #if __GNUC__ >= 2
336 re_token_type_t type : 8;
337 #else
338 re_token_type_t type;
339 #endif
340 unsigned int constraint : 10; /* context constraint */
341 unsigned int duplicated : 1;
342 unsigned int opt_subexp : 1;
343 #ifdef RE_ENABLE_I18N
344 unsigned int accept_mb : 1;
345 /* These 2 bits can be moved into the union if needed (e.g. if running out
346 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
347 unsigned int mb_partial : 1;
348 #endif
349 unsigned int word_char : 1;
350 } re_token_t;
351
352 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
353
354 struct re_string_t
355 {
356 /* Indicate the raw buffer which is the original string passed as an
357 argument of regexec(), re_search(), etc.. */
358 const unsigned char *raw_mbs;
359 /* Store the multibyte string. In case of "case insensitive mode" like
360 REG_ICASE, upper cases of the string are stored, otherwise MBS points
361 the same address that RAW_MBS points. */
362 unsigned char *mbs;
363 #ifdef RE_ENABLE_I18N
364 /* Store the wide character string which is corresponding to MBS. */
365 wint_t *wcs;
366 Idx *offsets;
367 mbstate_t cur_state;
368 #endif
369 /* Index in RAW_MBS. Each character mbs[i] corresponds to
370 raw_mbs[raw_mbs_idx + i]. */
371 Idx raw_mbs_idx;
372 /* The length of the valid characters in the buffers. */
373 Idx valid_len;
374 /* The corresponding number of bytes in raw_mbs array. */
375 Idx valid_raw_len;
376 /* The length of the buffers MBS and WCS. */
377 Idx bufs_len;
378 /* The index in MBS, which is updated by re_string_fetch_byte. */
379 Idx cur_idx;
380 /* length of RAW_MBS array. */
381 Idx raw_len;
382 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
383 Idx len;
384 /* End of the buffer may be shorter than its length in the cases such
385 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
386 instead of LEN. */
387 Idx raw_stop;
388 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
389 Idx stop;
390
391 /* The context of mbs[0]. We store the context independently, since
392 the context of mbs[0] may be different from raw_mbs[0], which is
393 the beginning of the input string. */
394 unsigned int tip_context;
395 /* The translation passed as a part of an argument of re_compile_pattern. */
396 unsigned REG_TRANSLATE_TYPE trans;
397 /* Copy of re_dfa_t's word_char. */
398 re_const_bitset_ptr_t word_char;
399 /* true if REG_ICASE. */
400 unsigned char icase;
401 unsigned char is_utf8;
402 unsigned char map_notascii;
403 unsigned char mbs_allocated;
404 unsigned char offsets_needed;
405 unsigned char newline_anchor;
406 unsigned char word_ops_used;
407 int mb_cur_max;
408 };
409 typedef struct re_string_t re_string_t;
410
411
412 struct re_dfa_t;
413 typedef struct re_dfa_t re_dfa_t;
414
415 #ifndef _LIBC
416 # ifdef __i386__
417 # define internal_function __attribute ((regparm (3), stdcall))
418 # else
419 # define internal_function
420 # endif
421 #endif
422
423 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
424 Idx new_buf_len)
425 internal_function;
426 #ifdef RE_ENABLE_I18N
427 static void build_wcs_buffer (re_string_t *pstr) internal_function;
428 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
429 internal_function;
430 #endif /* RE_ENABLE_I18N */
431 static void build_upper_buffer (re_string_t *pstr) internal_function;
432 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
433 static unsigned int re_string_context_at (const re_string_t *input,
434 Idx idx, int eflags)
435 internal_function __attribute ((pure));
436
437 #define re_string_peek_byte(pstr, offset) \
438 ((pstr)->mbs[(pstr)->cur_idx + offset])
439 #define re_string_fetch_byte(pstr) \
440 ((pstr)->mbs[(pstr)->cur_idx++])
441 #define re_string_first_byte(pstr, idx) \
442 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
443 #define re_string_is_single_byte_char(pstr, idx) \
444 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
445 || (pstr)->wcs[(idx) + 1] != WEOF))
446 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
447 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
448 #define re_string_get_buffer(pstr) ((pstr)->mbs)
449 #define re_string_length(pstr) ((pstr)->len)
450 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
451 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
452 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
453
454 #ifndef __NetBSD__
455 #include <alloca.h>
456 #endif
457
458 #ifndef _LIBC
459 # if HAVE_ALLOCA
460 /* The OS usually guarantees only one guard page at the bottom of the stack,
461 and a page size can be as small as 4096 bytes. So we cannot safely
462 allocate anything larger than 4096 bytes. Also care for the possibility
463 of a few compiler-allocated temporary stack slots. */
464 # define __libc_use_alloca(n) ((n) < 4032)
465 # else
466 /* alloca is implemented with malloc, so just use malloc. */
467 # define __libc_use_alloca(n) 0
468 # endif
469 #endif
470
471 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
472 #define re_xmalloc(t,n) ((t *) re_xnmalloc (n, sizeof (t)))
473 #define re_calloc(t,n) ((t *) calloc (n, sizeof (t)))
474 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
475 #define re_xrealloc(p,t,n) ((t *) re_xnrealloc (p, n, sizeof (t)))
476 #define re_x2realloc(p,t,pn) ((t *) re_x2nrealloc (p, pn, sizeof (t)))
477 #define re_free(p) free (p)
478
479 #ifndef SIZE_MAX
480 # define SIZE_MAX ((size_t) -1)
481 #endif
482
483 /* Return true if an array of N objects, each of size S, cannot exist
484 due to size arithmetic overflow. S must be nonzero. */
485 static inline bool
re_alloc_oversized(size_t n,size_t s)486 re_alloc_oversized (size_t n, size_t s)
487 {
488 return BE (SIZE_MAX / s < n, 0);
489 }
490
491 /* Return true if an array of (2 * N + 1) objects, each of size S,
492 cannot exist due to size arithmetic overflow. S must be nonzero. */
493 static inline bool
re_x2alloc_oversized(size_t n,size_t s)494 re_x2alloc_oversized (size_t n, size_t s)
495 {
496 return BE ((SIZE_MAX / s - 1) / 2 < n, 0);
497 }
498
499 /* Allocate an array of N objects, each with S bytes of memory,
500 dynamically, with error checking. S must be nonzero. */
501 static inline void *
re_xnmalloc(size_t n,size_t s)502 re_xnmalloc (size_t n, size_t s)
503 {
504 return re_alloc_oversized (n, s) ? NULL : malloc (n * s);
505 }
506
507 /* Change the size of an allocated block of memory P to an array of N
508 objects each of S bytes, with error checking. S must be nonzero. */
509 static inline void *
re_xnrealloc(void * p,size_t n,size_t s)510 re_xnrealloc (void *p, size_t n, size_t s)
511 {
512 return re_alloc_oversized (n, s) ? NULL : realloc (p, n * s);
513 }
514
515 /* Reallocate a block of memory P to an array of (2 * (*PN) + 1)
516 objects each of S bytes, with error checking. S must be nonzero.
517 If the allocation is successful, set *PN to the new allocation
518 count and return the resulting pointer. Otherwise, return
519 NULL. */
520 static inline void *
re_x2nrealloc(void * p,size_t * pn,size_t s)521 re_x2nrealloc (void *p, size_t *pn, size_t s)
522 {
523 if (re_x2alloc_oversized (*pn, s))
524 return NULL;
525 else
526 {
527 /* Add 1 in case *PN is zero. */
528 size_t n1 = 2 * *pn + 1;
529 p = realloc (p, n1 * s);
530 if (BE (p != NULL, 1))
531 *pn = n1;
532 return p;
533 }
534 }
535
536 struct bin_tree_t
537 {
538 struct bin_tree_t *parent;
539 struct bin_tree_t *left;
540 struct bin_tree_t *right;
541 struct bin_tree_t *first;
542 struct bin_tree_t *next;
543
544 re_token_t token;
545
546 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
547 Otherwise `type' indicate the type of this node. */
548 Idx node_idx;
549 };
550 typedef struct bin_tree_t bin_tree_t;
551
552 #define BIN_TREE_STORAGE_SIZE \
553 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
554
555 struct bin_tree_storage_t
556 {
557 struct bin_tree_storage_t *next;
558 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
559 };
560 typedef struct bin_tree_storage_t bin_tree_storage_t;
561
562 #define CONTEXT_WORD 1
563 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
564 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
565 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
566
567 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
568 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
569 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
570 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
571 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
572
573 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
574 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
575 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
576 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
577
578 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
579 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
580 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
581 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
582 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
583
584 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
585 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
586 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
587 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
588 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
589
590 struct re_dfastate_t
591 {
592 re_hashval_t hash;
593 re_node_set nodes;
594 re_node_set non_eps_nodes;
595 re_node_set inveclosure;
596 re_node_set *entrance_nodes;
597 struct re_dfastate_t **trtable, **word_trtable;
598 unsigned int context : 4;
599 unsigned int halt : 1;
600 /* If this state can accept `multi byte'.
601 Note that we refer to multibyte characters, and multi character
602 collating elements as `multi byte'. */
603 unsigned int accept_mb : 1;
604 /* If this state has backreference node(s). */
605 unsigned int has_backref : 1;
606 unsigned int has_constraint : 1;
607 };
608 typedef struct re_dfastate_t re_dfastate_t;
609
610 struct re_state_table_entry
611 {
612 Idx num;
613 Idx alloc;
614 re_dfastate_t **array;
615 };
616
617 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
618
619 typedef struct
620 {
621 Idx next_idx;
622 Idx alloc;
623 re_dfastate_t **array;
624 } state_array_t;
625
626 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
627
628 typedef struct
629 {
630 Idx node;
631 Idx str_idx; /* The position NODE match at. */
632 state_array_t path;
633 } re_sub_match_last_t;
634
635 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
636 And information about the node, whose type is OP_CLOSE_SUBEXP,
637 corresponding to NODE is stored in LASTS. */
638
639 typedef struct
640 {
641 Idx str_idx;
642 Idx node;
643 state_array_t *path;
644 Idx alasts; /* Allocation size of LASTS. */
645 Idx nlasts; /* The number of LASTS. */
646 re_sub_match_last_t **lasts;
647 } re_sub_match_top_t;
648
649 struct re_backref_cache_entry
650 {
651 Idx node;
652 Idx str_idx;
653 Idx subexp_from;
654 Idx subexp_to;
655 char more;
656 char unused;
657 unsigned short int eps_reachable_subexps_map;
658 };
659
660 typedef struct
661 {
662 /* The string object corresponding to the input string. */
663 re_string_t input;
664 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
665 re_dfa_t *const dfa;
666 #else
667 re_dfa_t *dfa;
668 #endif
669 /* EFLAGS of the argument of regexec. */
670 int eflags;
671 /* Where the matching ends. */
672 Idx match_last;
673 Idx last_node;
674 /* The state log used by the matcher. */
675 re_dfastate_t **state_log;
676 Idx state_log_top;
677 /* Back reference cache. */
678 Idx nbkref_ents;
679 Idx abkref_ents;
680 struct re_backref_cache_entry *bkref_ents;
681 int max_mb_elem_len;
682 Idx nsub_tops;
683 Idx asub_tops;
684 re_sub_match_top_t **sub_tops;
685 } re_match_context_t;
686
687 typedef struct
688 {
689 re_dfastate_t **sifted_states;
690 re_dfastate_t **limited_states;
691 Idx last_node;
692 Idx last_str_idx;
693 re_node_set limits;
694 } re_sift_context_t;
695
696 struct re_fail_stack_ent_t
697 {
698 Idx idx;
699 Idx node;
700 regmatch_t *regs;
701 re_node_set eps_via_nodes;
702 };
703
704 struct re_fail_stack_t
705 {
706 Idx num;
707 Idx alloc;
708 struct re_fail_stack_ent_t *stack;
709 };
710
711 struct re_dfa_t
712 {
713 re_token_t *nodes;
714 Idx nodes_alloc;
715 Idx nodes_len;
716 Idx *nexts;
717 Idx *org_indices;
718 re_node_set *edests;
719 re_node_set *eclosures;
720 re_node_set *inveclosures;
721 struct re_state_table_entry *state_table;
722 re_dfastate_t *init_state;
723 re_dfastate_t *init_state_word;
724 re_dfastate_t *init_state_nl;
725 re_dfastate_t *init_state_begbuf;
726 bin_tree_t *str_tree;
727 bin_tree_storage_t *str_tree_storage;
728 re_bitset_ptr_t sb_char;
729 int str_tree_storage_idx;
730
731 /* number of subexpressions `re_nsub' is in regex_t. */
732 re_hashval_t state_hash_mask;
733 Idx init_node;
734 Idx nbackref; /* The number of backreference in this dfa. */
735
736 /* Bitmap expressing which backreference is used. */
737 bitset_word used_bkref_map;
738 bitset_word completed_bkref_map;
739
740 unsigned int has_plural_match : 1;
741 /* If this dfa has "multibyte node", which is a backreference or
742 a node which can accept multibyte character or multi character
743 collating element. */
744 unsigned int has_mb_node : 1;
745 unsigned int is_utf8 : 1;
746 unsigned int map_notascii : 1;
747 unsigned int word_ops_used : 1;
748 int mb_cur_max;
749 bitset word_char;
750 reg_syntax_t syntax;
751 Idx *subexp_map;
752 #ifdef DEBUG
753 char* re_str;
754 #endif
755 __libc_lock_define (, lock)
756 };
757
758 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
759 #define re_node_set_remove(set,id) \
760 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
761 #define re_node_set_empty(p) ((p)->nelem = 0)
762 #define re_node_set_free(set) re_free ((set)->elems)
763
764 static void free_state (re_dfastate_t *state) internal_function;
765
766
767 typedef enum
768 {
769 SB_CHAR,
770 MB_CHAR,
771 EQUIV_CLASS,
772 COLL_SYM,
773 CHAR_CLASS
774 } bracket_elem_type;
775
776 typedef struct
777 {
778 bracket_elem_type type;
779 union
780 {
781 unsigned char ch;
782 unsigned char *name;
783 wchar_t wch;
784 } opr;
785 } bracket_elem_t;
786
787
788 /* Inline functions for bitset operation. */
789
790 static inline void
bitset_set(bitset set,Idx i)791 bitset_set (bitset set, Idx i)
792 {
793 set[i / BITSET_WORD_BITS] |= (bitset_word) 1 << i % BITSET_WORD_BITS;
794 }
795
796 static inline void
bitset_clear(bitset set,Idx i)797 bitset_clear (bitset set, Idx i)
798 {
799 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word) 1 << i % BITSET_WORD_BITS);
800 }
801
802 static inline bool
bitset_contain(const bitset set,Idx i)803 bitset_contain (const bitset set, Idx i)
804 {
805 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
806 }
807
808 static inline void
bitset_empty(bitset set)809 bitset_empty (bitset set)
810 {
811 memset (set, 0, sizeof (bitset));
812 }
813
814 static inline void
bitset_set_all(bitset set)815 bitset_set_all (bitset set)
816 {
817 memset (set, -1, sizeof (bitset_word) * (SBC_MAX / BITSET_WORD_BITS));
818 if (SBC_MAX % BITSET_WORD_BITS != 0)
819 set[BITSET_WORDS - 1] =
820 ((bitset_word) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
821 }
822
823 static inline void
bitset_copy(bitset dest,const bitset src)824 bitset_copy (bitset dest, const bitset src)
825 {
826 memcpy (dest, src, sizeof (bitset));
827 }
828
829 static inline void
bitset_not(bitset set)830 bitset_not (bitset set)
831 {
832 int i;
833 for (i = 0; i < SBC_MAX / BITSET_WORD_BITS; ++i)
834 set[i] = ~set[i];
835 if (SBC_MAX % BITSET_WORD_BITS != 0)
836 set[BITSET_WORDS - 1] =
837 ((((bitset_word) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
838 & ~set[BITSET_WORDS - 1]);
839 }
840
841 static inline void
bitset_merge(bitset dest,const bitset src)842 bitset_merge (bitset dest, const bitset src)
843 {
844 int i;
845 for (i = 0; i < BITSET_WORDS; ++i)
846 dest[i] |= src[i];
847 }
848
849 static inline void
bitset_mask(bitset dest,const bitset src)850 bitset_mask (bitset dest, const bitset src)
851 {
852 int i;
853 for (i = 0; i < BITSET_WORDS; ++i)
854 dest[i] &= src[i];
855 }
856
857 #if defined RE_ENABLE_I18N
858 /* Inline functions for re_string. */
859 static inline int
internal_function(pure)860 internal_function __attribute ((pure))
861 re_string_char_size_at (const re_string_t *pstr, Idx idx)
862 {
863 int byte_idx;
864 if (pstr->mb_cur_max == 1)
865 return 1;
866 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
867 if (pstr->wcs[idx + byte_idx] != WEOF)
868 break;
869 return byte_idx;
870 }
871
872 static inline wint_t
internal_function(pure)873 internal_function __attribute ((pure))
874 re_string_wchar_at (const re_string_t *pstr, Idx idx)
875 {
876 if (pstr->mb_cur_max == 1)
877 return (wint_t) pstr->mbs[idx];
878 return (wint_t) pstr->wcs[idx];
879 }
880
881 static int
internal_function(pure)882 internal_function __attribute ((pure))
883 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
884 {
885 #ifdef _LIBC
886 const unsigned char *p, *extra;
887 const int32_t *table, *indirect;
888 int32_t tmp;
889 # include <locale/weight.h>
890 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
891
892 if (nrules != 0)
893 {
894 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
895 extra = (const unsigned char *)
896 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
897 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
898 _NL_COLLATE_INDIRECTMB);
899 p = pstr->mbs + idx;
900 tmp = findidx (&p);
901 return p - pstr->mbs - idx;
902 }
903 else
904 #endif /* _LIBC */
905 return 1;
906 }
907 #endif /* RE_ENABLE_I18N */
908
909 #endif /* _REGEX_INTERNAL_H */
910