1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002-2022 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 The GNU C Library is free software; you can redistribute it and/or
7 modify it under the terms of the GNU Lesser General Public
8 License as published by the Free Software Foundation; either
9 version 2.1 of the License, or (at your option) any later version.
10
11 The GNU C Library 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 GNU
14 Lesser General Public License for more details.
15
16 You should have received a copy of the GNU Lesser General Public
17 License along with the GNU C Library; if not, see
18 <https://www.gnu.org/licenses/>. */
19
20 #ifndef _REGEX_INTERNAL_H
21 #define _REGEX_INTERNAL_H 1
22
23 #include <ctype.h>
24 #include <stdio.h>
25 #include <stdlib.h>
26 #include <string.h>
27
28 #include <langinfo.h>
29 #include <locale.h>
30 #include <wchar.h>
31 #include <wctype.h>
32 #include <stdbool.h>
33 #include <stdint.h>
34
35 #ifndef _LIBC
36 # include <dynarray.h>
37 #endif
38
39 #include <intprops.h>
40 #include <verify.h>
41
42 #if defined DEBUG && DEBUG != 0
43 # include <assert.h>
44 # define DEBUG_ASSERT(x) assert (x)
45 #else
46 # define DEBUG_ASSERT(x) assume (x)
47 #endif
48
49 #ifdef _LIBC
50 # include <libc-lock.h>
51 # define lock_define(name) __libc_lock_define (, name)
52 # define lock_init(lock) (__libc_lock_init (lock), 0)
53 # define lock_fini(lock) ((void) 0)
54 # define lock_lock(lock) __libc_lock_lock (lock)
55 # define lock_unlock(lock) __libc_lock_unlock (lock)
56 #elif defined GNULIB_LOCK && !defined GNULIB_REGEX_SINGLE_THREAD
57 # include "glthread/lock.h"
58 # define lock_define(name) gl_lock_define (, name)
59 # define lock_init(lock) glthread_lock_init (&(lock))
60 # define lock_fini(lock) glthread_lock_destroy (&(lock))
61 # define lock_lock(lock) glthread_lock_lock (&(lock))
62 # define lock_unlock(lock) glthread_lock_unlock (&(lock))
63 #elif defined GNULIB_PTHREAD && !defined GNULIB_REGEX_SINGLE_THREAD
64 # include <pthread.h>
65 # define lock_define(name) pthread_mutex_t name;
66 # define lock_init(lock) pthread_mutex_init (&(lock), 0)
67 # define lock_fini(lock) pthread_mutex_destroy (&(lock))
68 # define lock_lock(lock) pthread_mutex_lock (&(lock))
69 # define lock_unlock(lock) pthread_mutex_unlock (&(lock))
70 #else
71 # define lock_define(name)
72 # define lock_init(lock) 0
73 # define lock_fini(lock) ((void) 0)
74 /* The 'dfa' avoids an "unused variable 'dfa'" warning from GCC. */
75 # define lock_lock(lock) ((void) dfa)
76 # define lock_unlock(lock) ((void) 0)
77 #endif
78
79 /* In case that the system doesn't have isblank(). */
80 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
81 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
82 #endif
83
84 /* regex code assumes isascii has its usual numeric meaning,
85 even if the portable character set uses EBCDIC encoding,
86 and even if wint_t is wider than int. */
87 #ifndef _LIBC
88 # undef isascii
89 # define isascii(c) (((c) & ~0x7f) == 0)
90 #endif
91
92 #ifdef _LIBC
93 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
94 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
95 # include <locale/localeinfo.h>
96 # include <locale/coll-lookup.h>
97 # endif
98 #endif
99
100 /* This is for other GNU distributions with internationalized messages. */
101 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
102 # include <libintl.h>
103 # ifdef _LIBC
104 # undef gettext
105 # define gettext(msgid) \
106 __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
107 # endif
108 #else
109 # undef gettext
110 # define gettext(msgid) (msgid)
111 #endif
112
113 #ifndef gettext_noop
114 /* This define is so xgettext can find the internationalizable
115 strings. */
116 # define gettext_noop(String) String
117 #endif
118
119 /* Number of ASCII characters. */
120 #define ASCII_CHARS 0x80
121
122 /* Number of single byte characters. */
123 #define SBC_MAX (UCHAR_MAX + 1)
124
125 #define COLL_ELEM_LEN_MAX 8
126
127 /* The character which represents newline. */
128 #define NEWLINE_CHAR '\n'
129 #define WIDE_NEWLINE_CHAR L'\n'
130
131 /* Rename to standard API for using out of glibc. */
132 #ifndef _LIBC
133 # undef __wctype
134 # undef __iswalnum
135 # undef __iswctype
136 # undef __towlower
137 # undef __towupper
138 # define __wctype wctype
139 # define __iswalnum iswalnum
140 # define __iswctype iswctype
141 # define __towlower towlower
142 # define __towupper towupper
143 # define __btowc btowc
144 # define __mbrtowc mbrtowc
145 # define __wcrtomb wcrtomb
146 # define __regfree regfree
147 #endif /* not _LIBC */
148
149 /* Types related to integers. Unless protected by #ifdef _LIBC, the
150 regex code should avoid exact-width types like int32_t and uint64_t
151 as some non-GCC platforms lack them, an issue when this code is
152 used in Gnulib. */
153
154 #ifndef SSIZE_MAX
155 # define SSIZE_MAX ((ssize_t) (SIZE_MAX / 2))
156 #endif
157 #ifndef ULONG_WIDTH
158 # define ULONG_WIDTH REGEX_UINTEGER_WIDTH (ULONG_MAX)
159 /* The number of usable bits in an unsigned integer type with maximum
160 value MAX, as an int expression suitable in #if. Cover all known
161 practical hosts. This implementation exploits the fact that MAX is
162 1 less than a power of 2, and merely counts the number of 1 bits in
163 MAX; "COBn" means "count the number of 1 bits in the low-order n bits". */
164 # define REGEX_UINTEGER_WIDTH(max) REGEX_COB128 (max)
165 # define REGEX_COB128(n) (REGEX_COB64 ((n) >> 31 >> 31 >> 2) + REGEX_COB64 (n))
166 # define REGEX_COB64(n) (REGEX_COB32 ((n) >> 31 >> 1) + REGEX_COB32 (n))
167 # define REGEX_COB32(n) (REGEX_COB16 ((n) >> 16) + REGEX_COB16 (n))
168 # define REGEX_COB16(n) (REGEX_COB8 ((n) >> 8) + REGEX_COB8 (n))
169 # define REGEX_COB8(n) (REGEX_COB4 ((n) >> 4) + REGEX_COB4 (n))
170 # define REGEX_COB4(n) (!!((n) & 8) + !!((n) & 4) + !!((n) & 2) + ((n) & 1))
171 # if ULONG_MAX / 2 + 1 != 1ul << (ULONG_WIDTH - 1)
172 # error "ULONG_MAX out of range"
173 # endif
174 #endif
175
176 /* The type of indexes into strings. This is signed, not size_t,
177 since the API requires indexes to fit in regoff_t anyway, and using
178 signed integers makes the code a bit smaller and presumably faster.
179 The traditional GNU regex implementation uses int for indexes.
180 The POSIX-compatible implementation uses a possibly-wider type.
181 The name 'Idx' is three letters to minimize the hassle of
182 reindenting a lot of regex code that formerly used 'int'. */
183 typedef regoff_t Idx;
184 #ifdef _REGEX_LARGE_OFFSETS
185 # define IDX_MAX SSIZE_MAX
186 #else
187 # define IDX_MAX INT_MAX
188 #endif
189
190 /* A hash value, suitable for computing hash tables. */
191 typedef __re_size_t re_hashval_t;
192
193 /* An integer used to represent a set of bits. It must be unsigned,
194 and must be at least as wide as unsigned int. */
195 typedef unsigned long int bitset_word_t;
196 /* All bits set in a bitset_word_t. */
197 #define BITSET_WORD_MAX ULONG_MAX
198 /* Number of bits in a bitset_word_t. */
199 #define BITSET_WORD_BITS ULONG_WIDTH
200
201 /* Number of bitset_word_t values in a bitset_t. */
202 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
203
204 typedef bitset_word_t bitset_t[BITSET_WORDS];
205 typedef bitset_word_t *re_bitset_ptr_t;
206 typedef const bitset_word_t *re_const_bitset_ptr_t;
207
208 #define PREV_WORD_CONSTRAINT 0x0001
209 #define PREV_NOTWORD_CONSTRAINT 0x0002
210 #define NEXT_WORD_CONSTRAINT 0x0004
211 #define NEXT_NOTWORD_CONSTRAINT 0x0008
212 #define PREV_NEWLINE_CONSTRAINT 0x0010
213 #define NEXT_NEWLINE_CONSTRAINT 0x0020
214 #define PREV_BEGBUF_CONSTRAINT 0x0040
215 #define NEXT_ENDBUF_CONSTRAINT 0x0080
216 #define WORD_DELIM_CONSTRAINT 0x0100
217 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
218
219 typedef enum
220 {
221 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
222 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
223 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
224 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
225 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
226 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
227 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
228 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
229 WORD_DELIM = WORD_DELIM_CONSTRAINT,
230 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
231 } re_context_type;
232
233 typedef struct
234 {
235 Idx alloc;
236 Idx nelem;
237 Idx *elems;
238 } re_node_set;
239
240 typedef enum
241 {
242 NON_TYPE = 0,
243
244 /* Node type, These are used by token, node, tree. */
245 CHARACTER = 1,
246 END_OF_RE = 2,
247 SIMPLE_BRACKET = 3,
248 OP_BACK_REF = 4,
249 OP_PERIOD = 5,
250 COMPLEX_BRACKET = 6,
251 OP_UTF8_PERIOD = 7,
252
253 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
254 when the debugger shows values of this enum type. */
255 #define EPSILON_BIT 8
256 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
257 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
258 OP_ALT = EPSILON_BIT | 2,
259 OP_DUP_ASTERISK = EPSILON_BIT | 3,
260 ANCHOR = EPSILON_BIT | 4,
261
262 /* Tree type, these are used only by tree. */
263 CONCAT = 16,
264 SUBEXP = 17,
265
266 /* Token type, these are used only by token. */
267 OP_DUP_PLUS = 18,
268 OP_DUP_QUESTION,
269 OP_OPEN_BRACKET,
270 OP_CLOSE_BRACKET,
271 OP_CHARSET_RANGE,
272 OP_OPEN_DUP_NUM,
273 OP_CLOSE_DUP_NUM,
274 OP_NON_MATCH_LIST,
275 OP_OPEN_COLL_ELEM,
276 OP_CLOSE_COLL_ELEM,
277 OP_OPEN_EQUIV_CLASS,
278 OP_CLOSE_EQUIV_CLASS,
279 OP_OPEN_CHAR_CLASS,
280 OP_CLOSE_CHAR_CLASS,
281 OP_WORD,
282 OP_NOTWORD,
283 OP_SPACE,
284 OP_NOTSPACE,
285 BACK_SLASH
286
287 } re_token_type_t;
288
289 typedef struct
290 {
291 /* Multibyte characters. */
292 wchar_t *mbchars;
293
294 #ifdef _LIBC
295 /* Collating symbols. */
296 int32_t *coll_syms;
297 #endif
298
299 #ifdef _LIBC
300 /* Equivalence classes. */
301 int32_t *equiv_classes;
302 #endif
303
304 /* Range expressions. */
305 #ifdef _LIBC
306 uint32_t *range_starts;
307 uint32_t *range_ends;
308 #else
309 wchar_t *range_starts;
310 wchar_t *range_ends;
311 #endif
312
313 /* Character classes. */
314 wctype_t *char_classes;
315
316 /* If this character set is the non-matching list. */
317 unsigned int non_match : 1;
318
319 /* # of multibyte characters. */
320 Idx nmbchars;
321
322 /* # of collating symbols. */
323 Idx ncoll_syms;
324
325 /* # of equivalence classes. */
326 Idx nequiv_classes;
327
328 /* # of range expressions. */
329 Idx nranges;
330
331 /* # of character classes. */
332 Idx nchar_classes;
333 } re_charset_t;
334
335 typedef struct
336 {
337 union
338 {
339 unsigned char c; /* for CHARACTER */
340 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
341 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
342 Idx idx; /* for BACK_REF */
343 re_context_type ctx_type; /* for ANCHOR */
344 } opr;
345 #if (__GNUC__ >= 2 || defined __clang__) && !defined __STRICT_ANSI__
346 re_token_type_t type : 8;
347 #else
348 re_token_type_t type;
349 #endif
350 unsigned int constraint : 10; /* context constraint */
351 unsigned int duplicated : 1;
352 unsigned int opt_subexp : 1;
353 unsigned int accept_mb : 1;
354 /* These 2 bits can be moved into the union if needed (e.g. if running out
355 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
356 unsigned int mb_partial : 1;
357 unsigned int word_char : 1;
358 } re_token_t;
359
360 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
361
362 struct re_string_t
363 {
364 /* Indicate the raw buffer which is the original string passed as an
365 argument of regexec(), re_search(), etc.. */
366 const unsigned char *raw_mbs;
367 /* Store the multibyte string. In case of "case insensitive mode" like
368 REG_ICASE, upper cases of the string are stored, otherwise MBS points
369 the same address that RAW_MBS points. */
370 unsigned char *mbs;
371 /* Store the wide character string which is corresponding to MBS. */
372 wint_t *wcs;
373 Idx *offsets;
374 mbstate_t cur_state;
375 /* Index in RAW_MBS. Each character mbs[i] corresponds to
376 raw_mbs[raw_mbs_idx + i]. */
377 Idx raw_mbs_idx;
378 /* The length of the valid characters in the buffers. */
379 Idx valid_len;
380 /* The corresponding number of bytes in raw_mbs array. */
381 Idx valid_raw_len;
382 /* The length of the buffers MBS and WCS. */
383 Idx bufs_len;
384 /* The index in MBS, which is updated by re_string_fetch_byte. */
385 Idx cur_idx;
386 /* length of RAW_MBS array. */
387 Idx raw_len;
388 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
389 Idx len;
390 /* End of the buffer may be shorter than its length in the cases such
391 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
392 instead of LEN. */
393 Idx raw_stop;
394 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
395 Idx stop;
396
397 /* The context of mbs[0]. We store the context independently, since
398 the context of mbs[0] may be different from raw_mbs[0], which is
399 the beginning of the input string. */
400 unsigned int tip_context;
401 /* The translation passed as a part of an argument of re_compile_pattern. */
402 RE_TRANSLATE_TYPE trans;
403 /* Copy of re_dfa_t's word_char. */
404 re_const_bitset_ptr_t word_char;
405 /* true if REG_ICASE. */
406 unsigned char icase;
407 unsigned char is_utf8;
408 unsigned char map_notascii;
409 unsigned char mbs_allocated;
410 unsigned char offsets_needed;
411 unsigned char newline_anchor;
412 unsigned char word_ops_used;
413 int mb_cur_max;
414 };
415 typedef struct re_string_t re_string_t;
416
417
418 struct re_dfa_t;
419 typedef struct re_dfa_t re_dfa_t;
420
421 #ifndef _LIBC
422 # define IS_IN(libc) false
423 #endif
424
425 #define re_string_peek_byte(pstr, offset) \
426 ((pstr)->mbs[(pstr)->cur_idx + offset])
427 #define re_string_fetch_byte(pstr) \
428 ((pstr)->mbs[(pstr)->cur_idx++])
429 #define re_string_first_byte(pstr, idx) \
430 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
431 #define re_string_is_single_byte_char(pstr, idx) \
432 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
433 || (pstr)->wcs[(idx) + 1] != WEOF))
434 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
435 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
436 #define re_string_get_buffer(pstr) ((pstr)->mbs)
437 #define re_string_length(pstr) ((pstr)->len)
438 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
439 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
440 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
441
442 #ifdef _LIBC
443 # define MALLOC_0_IS_NONNULL 1
444 #elif !defined MALLOC_0_IS_NONNULL
445 # define MALLOC_0_IS_NONNULL 0
446 #endif
447
448 #ifndef MAX
449 # define MAX(a,b) ((a) < (b) ? (b) : (a))
450 #endif
451 #ifndef MIN
452 # define MIN(a,b) ((a) < (b) ? (a) : (b))
453 #endif
454
455 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
456 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
457 #define re_free(p) free (p)
458
459 struct bin_tree_t
460 {
461 struct bin_tree_t *parent;
462 struct bin_tree_t *left;
463 struct bin_tree_t *right;
464 struct bin_tree_t *first;
465 struct bin_tree_t *next;
466
467 re_token_t token;
468
469 /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
470 Otherwise 'type' indicate the type of this node. */
471 Idx node_idx;
472 };
473 typedef struct bin_tree_t bin_tree_t;
474
475 #define BIN_TREE_STORAGE_SIZE \
476 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
477
478 struct bin_tree_storage_t
479 {
480 struct bin_tree_storage_t *next;
481 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
482 };
483 typedef struct bin_tree_storage_t bin_tree_storage_t;
484
485 #define CONTEXT_WORD 1
486 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
487 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
488 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
489
490 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
491 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
492 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
493 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
494 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
495
496 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
497 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
498 #define IS_WIDE_WORD_CHAR(ch) (__iswalnum (ch) || (ch) == L'_')
499 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
500
501 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
502 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
503 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
504 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
505 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
506
507 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
508 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
509 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
510 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
511 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
512
513 struct re_dfastate_t
514 {
515 re_hashval_t hash;
516 re_node_set nodes;
517 re_node_set non_eps_nodes;
518 re_node_set inveclosure;
519 re_node_set *entrance_nodes;
520 struct re_dfastate_t **trtable, **word_trtable;
521 unsigned int context : 4;
522 unsigned int halt : 1;
523 /* If this state can accept "multi byte".
524 Note that we refer to multibyte characters, and multi character
525 collating elements as "multi byte". */
526 unsigned int accept_mb : 1;
527 /* If this state has backreference node(s). */
528 unsigned int has_backref : 1;
529 unsigned int has_constraint : 1;
530 };
531 typedef struct re_dfastate_t re_dfastate_t;
532
533 struct re_state_table_entry
534 {
535 Idx num;
536 Idx alloc;
537 re_dfastate_t **array;
538 };
539
540 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
541
542 typedef struct
543 {
544 Idx next_idx;
545 Idx alloc;
546 re_dfastate_t **array;
547 } state_array_t;
548
549 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
550
551 typedef struct
552 {
553 Idx node;
554 Idx str_idx; /* The position NODE match at. */
555 state_array_t path;
556 } re_sub_match_last_t;
557
558 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
559 And information about the node, whose type is OP_CLOSE_SUBEXP,
560 corresponding to NODE is stored in LASTS. */
561
562 typedef struct
563 {
564 Idx str_idx;
565 Idx node;
566 state_array_t *path;
567 Idx alasts; /* Allocation size of LASTS. */
568 Idx nlasts; /* The number of LASTS. */
569 re_sub_match_last_t **lasts;
570 } re_sub_match_top_t;
571
572 struct re_backref_cache_entry
573 {
574 Idx node;
575 Idx str_idx;
576 Idx subexp_from;
577 Idx subexp_to;
578 bitset_word_t eps_reachable_subexps_map;
579 char more;
580 };
581
582 typedef struct
583 {
584 /* The string object corresponding to the input string. */
585 re_string_t input;
586 const re_dfa_t *const dfa;
587 /* EFLAGS of the argument of regexec. */
588 int eflags;
589 /* Where the matching ends. */
590 Idx match_last;
591 Idx last_node;
592 /* The state log used by the matcher. */
593 re_dfastate_t **state_log;
594 Idx state_log_top;
595 /* Back reference cache. */
596 Idx nbkref_ents;
597 Idx abkref_ents;
598 struct re_backref_cache_entry *bkref_ents;
599 int max_mb_elem_len;
600 Idx nsub_tops;
601 Idx asub_tops;
602 re_sub_match_top_t **sub_tops;
603 } re_match_context_t;
604
605 typedef struct
606 {
607 re_dfastate_t **sifted_states;
608 re_dfastate_t **limited_states;
609 Idx last_node;
610 Idx last_str_idx;
611 re_node_set limits;
612 } re_sift_context_t;
613
614 struct re_fail_stack_ent_t
615 {
616 Idx idx;
617 Idx node;
618 regmatch_t *regs;
619 re_node_set eps_via_nodes;
620 };
621
622 struct re_fail_stack_t
623 {
624 Idx num;
625 Idx alloc;
626 struct re_fail_stack_ent_t *stack;
627 };
628
629 struct re_dfa_t
630 {
631 re_token_t *nodes;
632 size_t nodes_alloc;
633 size_t nodes_len;
634 Idx *nexts;
635 Idx *org_indices;
636 re_node_set *edests;
637 re_node_set *eclosures;
638 re_node_set *inveclosures;
639 struct re_state_table_entry *state_table;
640 re_dfastate_t *init_state;
641 re_dfastate_t *init_state_word;
642 re_dfastate_t *init_state_nl;
643 re_dfastate_t *init_state_begbuf;
644 bin_tree_t *str_tree;
645 bin_tree_storage_t *str_tree_storage;
646 re_bitset_ptr_t sb_char;
647 int str_tree_storage_idx;
648
649 /* number of subexpressions 're_nsub' is in regex_t. */
650 re_hashval_t state_hash_mask;
651 Idx init_node;
652 Idx nbackref; /* The number of backreference in this dfa. */
653
654 /* Bitmap expressing which backreference is used. */
655 bitset_word_t used_bkref_map;
656 bitset_word_t completed_bkref_map;
657
658 unsigned int has_plural_match : 1;
659 /* If this dfa has "multibyte node", which is a backreference or
660 a node which can accept multibyte character or multi character
661 collating element. */
662 unsigned int has_mb_node : 1;
663 unsigned int is_utf8 : 1;
664 unsigned int map_notascii : 1;
665 unsigned int word_ops_used : 1;
666 int mb_cur_max;
667 bitset_t word_char;
668 reg_syntax_t syntax;
669 Idx *subexp_map;
670 #ifdef DEBUG
671 char* re_str;
672 #endif
673 lock_define (lock)
674 };
675
676 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
677 #define re_node_set_remove(set,id) \
678 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
679 #define re_node_set_empty(p) ((p)->nelem = 0)
680 #define re_node_set_free(set) re_free ((set)->elems)
681
682
683 typedef enum
684 {
685 SB_CHAR,
686 MB_CHAR,
687 EQUIV_CLASS,
688 COLL_SYM,
689 CHAR_CLASS
690 } bracket_elem_type;
691
692 typedef struct
693 {
694 bracket_elem_type type;
695 union
696 {
697 unsigned char ch;
698 unsigned char *name;
699 wchar_t wch;
700 } opr;
701 } bracket_elem_t;
702
703
704 /* Functions for bitset_t operation. */
705
706 static inline void
707 bitset_set (bitset_t set, Idx i)
708 {
709 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
710 }
711
712 static inline void
713 bitset_clear (bitset_t set, Idx i)
714 {
715 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
716 }
717
718 static inline bool
719 bitset_contain (const bitset_t set, Idx i)
720 {
721 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
722 }
723
724 static inline void
725 bitset_empty (bitset_t set)
726 {
727 memset (set, '\0', sizeof (bitset_t));
728 }
729
730 static inline void
731 bitset_set_all (bitset_t set)
732 {
733 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
734 if (SBC_MAX % BITSET_WORD_BITS != 0)
735 set[BITSET_WORDS - 1] =
736 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
737 }
738
739 static inline void
740 bitset_copy (bitset_t dest, const bitset_t src)
741 {
742 memcpy (dest, src, sizeof (bitset_t));
743 }
744
745 static inline void
746 bitset_not (bitset_t set)
747 {
748 int bitset_i;
749 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
750 set[bitset_i] = ~set[bitset_i];
751 if (SBC_MAX % BITSET_WORD_BITS != 0)
752 set[BITSET_WORDS - 1] =
753 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
754 & ~set[BITSET_WORDS - 1]);
755 }
756
757 static inline void
758 bitset_merge (bitset_t dest, const bitset_t src)
759 {
760 int bitset_i;
761 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
762 dest[bitset_i] |= src[bitset_i];
763 }
764
765 static inline void
766 bitset_mask (bitset_t dest, const bitset_t src)
767 {
768 int bitset_i;
769 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
770 dest[bitset_i] &= src[bitset_i];
771 }
772
773 /* Functions for re_string. */
774 static int
775 __attribute__ ((pure, unused))
776 re_string_char_size_at (const re_string_t *pstr, Idx idx)
777 {
778 int byte_idx;
779 if (pstr->mb_cur_max == 1)
780 return 1;
781 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
782 if (pstr->wcs[idx + byte_idx] != WEOF)
783 break;
784 return byte_idx;
785 }
786
787 static wint_t
788 __attribute__ ((pure, unused))
789 re_string_wchar_at (const re_string_t *pstr, Idx idx)
790 {
791 if (pstr->mb_cur_max == 1)
792 return (wint_t) pstr->mbs[idx];
793 return (wint_t) pstr->wcs[idx];
794 }
795
796 #ifdef _LIBC
797 # include <locale/weight.h>
798 #endif
799
800 static int
801 __attribute__ ((pure, unused))
802 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
803 {
804 #ifdef _LIBC
805 const unsigned char *p, *extra;
806 const int32_t *table, *indirect;
807 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
808
809 if (nrules != 0)
810 {
811 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
812 extra = (const unsigned char *)
813 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
814 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
815 _NL_COLLATE_INDIRECTMB);
816 p = pstr->mbs + idx;
817 findidx (table, indirect, extra, &p, pstr->len - idx);
818 return p - pstr->mbs - idx;
819 }
820 #endif /* _LIBC */
821
822 return 1;
823 }
824
825 #ifdef _LIBC
826 # if __GNUC__ >= 7
827 # define FALLTHROUGH __attribute__ ((__fallthrough__))
828 # else
829 # define FALLTHROUGH ((void) 0)
830 # endif
831 #else
832 # include "attribute.h"
833 #endif
834
835 #endif /* _REGEX_INTERNAL_H */