1/* Extended regular expression matching and search library.
2 Copyright (C) 2002-2018 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 <http://www.gnu.org/licenses/>. */
19
20#ifndef _REGEX_INTERNAL_H
21#define _REGEX_INTERNAL_H 1
22
23#include <assert.h>
24#include <ctype.h>
25#include <stdio.h>
26#include <stdlib.h>
27#include <string.h>
28
29#if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET || defined _LIBC
30# include <langinfo.h>
31#endif
32#if defined HAVE_LOCALE_H || defined _LIBC
33# include <locale.h>
34#endif
35#if defined HAVE_WCHAR_H || defined _LIBC
36# include <wchar.h>
37#endif /* HAVE_WCHAR_H || _LIBC */
38#if defined HAVE_WCTYPE_H || defined _LIBC
39# include <wctype.h>
40#endif /* HAVE_WCTYPE_H || _LIBC */
41#if defined HAVE_STDBOOL_H || defined _LIBC
42# include <stdbool.h>
43#endif /* HAVE_STDBOOL_H || _LIBC */
44#if defined HAVE_STDINT_H || defined _LIBC
45# include <stdint.h>
46#endif /* HAVE_STDINT_H || _LIBC */
47#if defined _LIBC
48# include <libc-lock.h>
49#else
50# define __libc_lock_define(CLASS,NAME)
51# define __libc_lock_init(NAME) do { } while (0)
52# define __libc_lock_lock(NAME) do { } while (0)
53# define __libc_lock_unlock(NAME) do { } while (0)
54#endif
55
56/* In case that the system doesn't have isblank(). */
57#if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank
58# define isblank(ch) ((ch) == ' ' || (ch) == '\t')
59#endif
60
61#ifdef _LIBC
62# ifndef _RE_DEFINE_LOCALE_FUNCTIONS
63# define _RE_DEFINE_LOCALE_FUNCTIONS 1
64# include <locale/localeinfo.h>
65# include <locale/coll-lookup.h>
66# endif
67#endif
68
69/* This is for other GNU distributions with internationalized messages. */
70#if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
71# include <libintl.h>
72# ifdef _LIBC
73# undef gettext
74# define gettext(msgid) \
75 __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
76# endif
77#else
78# define gettext(msgid) (msgid)
79#endif
80
81#ifndef gettext_noop
82/* This define is so xgettext can find the internationalizable
83 strings. */
84# define gettext_noop(String) String
85#endif
86
87/* For loser systems without the definition. */
88#ifndef SIZE_MAX
89# define SIZE_MAX ((size_t) -1)
90#endif
91
92#if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE) || _LIBC
93# define RE_ENABLE_I18N
94#endif
95
96#if __GNUC__ >= 3
97# define BE(expr, val) __builtin_expect (expr, val)
98#else
99# define BE(expr, val) (expr)
100#endif
101
102/* Number of single byte character. */
103#define SBC_MAX 256
104
105#define COLL_ELEM_LEN_MAX 8
106
107/* The character which represents newline. */
108#define NEWLINE_CHAR '\n'
109#define WIDE_NEWLINE_CHAR L'\n'
110
111/* Rename to standard API for using out of glibc. */
112#ifndef _LIBC
113# define __wctype wctype
114# define __iswctype iswctype
115# define __btowc btowc
116# define __mbrtowc mbrtowc
117# define __mempcpy mempcpy
118# define __wcrtomb wcrtomb
119# define __regfree regfree
120# define attribute_hidden
121#endif /* not _LIBC */
122
123#if __GNUC__ < 3 + (__GNUC_MINOR__ < 1)
124# define __attribute__(arg)
125#endif
126
127extern const char __re_error_msgid[] attribute_hidden;
128extern const size_t __re_error_msgid_idx[] attribute_hidden;
129
130/* An integer used to represent a set of bits. It must be unsigned,
131 and must be at least as wide as unsigned int. */
132typedef unsigned long int bitset_word_t;
133/* All bits set in a bitset_word_t. */
134#define BITSET_WORD_MAX ULONG_MAX
135/* Number of bits in a bitset_word_t. */
136#define BITSET_WORD_BITS (sizeof (bitset_word_t) * CHAR_BIT)
137/* Number of bitset_word_t in a bit_set. */
138#define BITSET_WORDS (SBC_MAX / BITSET_WORD_BITS)
139typedef bitset_word_t bitset_t[BITSET_WORDS];
140typedef bitset_word_t *re_bitset_ptr_t;
141typedef const bitset_word_t *re_const_bitset_ptr_t;
142
143#define bitset_set(set,i) \
144 (set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS)
145#define bitset_clear(set,i) \
146 (set[i / BITSET_WORD_BITS] &= ~((bitset_word_t) 1 << i % BITSET_WORD_BITS))
147#define bitset_contain(set,i) \
148 (set[i / BITSET_WORD_BITS] & ((bitset_word_t) 1 << i % BITSET_WORD_BITS))
149#define bitset_empty(set) memset (set, '\0', sizeof (bitset_t))
150#define bitset_set_all(set) memset (set, '\xff', sizeof (bitset_t))
151#define bitset_copy(dest,src) memcpy (dest, src, sizeof (bitset_t))
152
153#define PREV_WORD_CONSTRAINT 0x0001
154#define PREV_NOTWORD_CONSTRAINT 0x0002
155#define NEXT_WORD_CONSTRAINT 0x0004
156#define NEXT_NOTWORD_CONSTRAINT 0x0008
157#define PREV_NEWLINE_CONSTRAINT 0x0010
158#define NEXT_NEWLINE_CONSTRAINT 0x0020
159#define PREV_BEGBUF_CONSTRAINT 0x0040
160#define NEXT_ENDBUF_CONSTRAINT 0x0080
161#define WORD_DELIM_CONSTRAINT 0x0100
162#define NOT_WORD_DELIM_CONSTRAINT 0x0200
163
164typedef enum
165{
166 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
167 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
168 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
169 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
170 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
171 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
172 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
173 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
174 WORD_DELIM = WORD_DELIM_CONSTRAINT,
175 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
176} re_context_type;
177
178typedef struct
179{
180 int alloc;
181 int nelem;
182 int *elems;
183} re_node_set;
184
185typedef enum
186{
187 NON_TYPE = 0,
188
189 /* Node type, These are used by token, node, tree. */
190 CHARACTER = 1,
191 END_OF_RE = 2,
192 SIMPLE_BRACKET = 3,
193 OP_BACK_REF = 4,
194 OP_PERIOD = 5,
195#ifdef RE_ENABLE_I18N
196 COMPLEX_BRACKET = 6,
197 OP_UTF8_PERIOD = 7,
198#endif /* RE_ENABLE_I18N */
199
200 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
201 when the debugger shows values of this enum type. */
202#define EPSILON_BIT 8
203 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
204 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
205 OP_ALT = EPSILON_BIT | 2,
206 OP_DUP_ASTERISK = EPSILON_BIT | 3,
207 ANCHOR = EPSILON_BIT | 4,
208
209 /* Tree type, these are used only by tree. */
210 CONCAT = 16,
211 SUBEXP = 17,
212
213 /* Token type, these are used only by token. */
214 OP_DUP_PLUS = 18,
215 OP_DUP_QUESTION,
216 OP_OPEN_BRACKET,
217 OP_CLOSE_BRACKET,
218 OP_CHARSET_RANGE,
219 OP_OPEN_DUP_NUM,
220 OP_CLOSE_DUP_NUM,
221 OP_NON_MATCH_LIST,
222 OP_OPEN_COLL_ELEM,
223 OP_CLOSE_COLL_ELEM,
224 OP_OPEN_EQUIV_CLASS,
225 OP_CLOSE_EQUIV_CLASS,
226 OP_OPEN_CHAR_CLASS,
227 OP_CLOSE_CHAR_CLASS,
228 OP_WORD,
229 OP_NOTWORD,
230 OP_SPACE,
231 OP_NOTSPACE,
232 BACK_SLASH
233
234} re_token_type_t;
235
236#ifdef RE_ENABLE_I18N
237typedef struct
238{
239 /* Multibyte characters. */
240 wchar_t *mbchars;
241
242 /* Collating symbols. */
243# ifdef _LIBC
244 int32_t *coll_syms;
245# endif
246
247 /* Equivalence classes. */
248# ifdef _LIBC
249 int32_t *equiv_classes;
250# endif
251
252 /* Range expressions. */
253# ifdef _LIBC
254 uint32_t *range_starts;
255 uint32_t *range_ends;
256# else /* not _LIBC */
257 wchar_t *range_starts;
258 wchar_t *range_ends;
259# endif /* not _LIBC */
260
261 /* Character classes. */
262 wctype_t *char_classes;
263
264 /* If this character set is the non-matching list. */
265 unsigned int non_match : 1;
266
267 /* # of multibyte characters. */
268 int nmbchars;
269
270 /* # of collating symbols. */
271 int ncoll_syms;
272
273 /* # of equivalence classes. */
274 int nequiv_classes;
275
276 /* # of range expressions. */
277 int nranges;
278
279 /* # of character classes. */
280 int nchar_classes;
281} re_charset_t;
282#endif /* RE_ENABLE_I18N */
283
284typedef struct
285{
286 union
287 {
288 unsigned char c; /* for CHARACTER */
289 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
290#ifdef RE_ENABLE_I18N
291 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
292#endif /* RE_ENABLE_I18N */
293 int idx; /* for BACK_REF */
294 re_context_type ctx_type; /* for ANCHOR */
295 } opr;
296#if __GNUC__ >= 2
297 re_token_type_t type : 8;
298#else
299 re_token_type_t type;
300#endif
301 unsigned int constraint : 10; /* context constraint */
302 unsigned int duplicated : 1;
303 unsigned int opt_subexp : 1;
304#ifdef RE_ENABLE_I18N
305 unsigned int accept_mb : 1;
306 /* These 2 bits can be moved into the union if needed (e.g. if running out
307 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
308 unsigned int mb_partial : 1;
309#endif
310 unsigned int word_char : 1;
311} re_token_t;
312
313#define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
314
315struct re_string_t
316{
317 /* Indicate the raw buffer which is the original string passed as an
318 argument of regexec(), re_search(), etc.. */
319 const unsigned char *raw_mbs;
320 /* Store the multibyte string. In case of "case insensitive mode" like
321 REG_ICASE, upper cases of the string are stored, otherwise MBS points
322 the same address that RAW_MBS points. */
323 unsigned char *mbs;
324#ifdef RE_ENABLE_I18N
325 /* Store the wide character string which is corresponding to MBS. */
326 wint_t *wcs;
327 int *offsets;
328 mbstate_t cur_state;
329#endif
330 /* Index in RAW_MBS. Each character mbs[i] corresponds to
331 raw_mbs[raw_mbs_idx + i]. */
332 int raw_mbs_idx;
333 /* The length of the valid characters in the buffers. */
334 int valid_len;
335 /* The corresponding number of bytes in raw_mbs array. */
336 int valid_raw_len;
337 /* The length of the buffers MBS and WCS. */
338 int bufs_len;
339 /* The index in MBS, which is updated by re_string_fetch_byte. */
340 int cur_idx;
341 /* length of RAW_MBS array. */
342 int raw_len;
343 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
344 int len;
345 /* End of the buffer may be shorter than its length in the cases such
346 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
347 instead of LEN. */
348 int raw_stop;
349 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
350 int stop;
351
352 /* The context of mbs[0]. We store the context independently, since
353 the context of mbs[0] may be different from raw_mbs[0], which is
354 the beginning of the input string. */
355 unsigned int tip_context;
356 /* The translation passed as a part of an argument of re_compile_pattern. */
357 RE_TRANSLATE_TYPE trans;
358 /* Copy of re_dfa_t's word_char. */
359 re_const_bitset_ptr_t word_char;
360 /* 1 if REG_ICASE. */
361 unsigned char icase;
362 unsigned char is_utf8;
363 unsigned char map_notascii;
364 unsigned char mbs_allocated;
365 unsigned char offsets_needed;
366 unsigned char newline_anchor;
367 unsigned char word_ops_used;
368 int mb_cur_max;
369};
370typedef struct re_string_t re_string_t;
371
372
373struct re_dfa_t;
374typedef struct re_dfa_t re_dfa_t;
375
376#if IS_IN (libc)
377static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
378 int new_buf_len);
379# ifdef RE_ENABLE_I18N
380static void build_wcs_buffer (re_string_t *pstr);
381static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr);
382# endif /* RE_ENABLE_I18N */
383static void build_upper_buffer (re_string_t *pstr);
384static void re_string_translate_buffer (re_string_t *pstr);
385static unsigned int re_string_context_at (const re_string_t *input, int idx,
386 int eflags) __attribute__ ((pure));
387#endif
388#define re_string_peek_byte(pstr, offset) \
389 ((pstr)->mbs[(pstr)->cur_idx + offset])
390#define re_string_fetch_byte(pstr) \
391 ((pstr)->mbs[(pstr)->cur_idx++])
392#define re_string_first_byte(pstr, idx) \
393 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
394#define re_string_is_single_byte_char(pstr, idx) \
395 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
396 || (pstr)->wcs[(idx) + 1] != WEOF))
397#define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
398#define re_string_cur_idx(pstr) ((pstr)->cur_idx)
399#define re_string_get_buffer(pstr) ((pstr)->mbs)
400#define re_string_length(pstr) ((pstr)->len)
401#define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
402#define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
403#define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
404
405#include <alloca.h>
406
407#ifndef _LIBC
408# if HAVE_ALLOCA
409/* The OS usually guarantees only one guard page at the bottom of the stack,
410 and a page size can be as small as 4096 bytes. So we cannot safely
411 allocate anything larger than 4096 bytes. Also care for the possibility
412 of a few compiler-allocated temporary stack slots. */
413# define __libc_use_alloca(n) ((n) < 4032)
414# else
415/* alloca is implemented with malloc, so just use malloc. */
416# define __libc_use_alloca(n) 0
417# endif
418#endif
419
420#define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
421#define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
422#define re_free(p) free (p)
423
424struct bin_tree_t
425{
426 struct bin_tree_t *parent;
427 struct bin_tree_t *left;
428 struct bin_tree_t *right;
429 struct bin_tree_t *first;
430 struct bin_tree_t *next;
431
432 re_token_t token;
433
434 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
435 Otherwise `type' indicate the type of this node. */
436 int node_idx;
437};
438typedef struct bin_tree_t bin_tree_t;
439
440#define BIN_TREE_STORAGE_SIZE \
441 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
442
443struct bin_tree_storage_t
444{
445 struct bin_tree_storage_t *next;
446 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
447};
448typedef struct bin_tree_storage_t bin_tree_storage_t;
449
450#define CONTEXT_WORD 1
451#define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
452#define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
453#define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
454
455#define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
456#define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
457#define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
458#define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
459#define IS_ORDINARY_CONTEXT(c) ((c) == 0)
460
461#define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
462#define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
463#define IS_WIDE_WORD_CHAR(ch) (__iswalnum (ch) || (ch) == L'_')
464#define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
465
466#define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
467 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
468 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
469 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
470 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
471
472#define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
473 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
474 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
475 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
476 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
477
478struct re_dfastate_t
479{
480 unsigned int hash;
481 re_node_set nodes;
482 re_node_set non_eps_nodes;
483 re_node_set inveclosure;
484 re_node_set *entrance_nodes;
485 struct re_dfastate_t **trtable, **word_trtable;
486 unsigned int context : 4;
487 unsigned int halt : 1;
488 /* If this state can accept `multi byte'.
489 Note that we refer to multibyte characters, and multi character
490 collating elements as `multi byte'. */
491 unsigned int accept_mb : 1;
492 /* If this state has backreference node(s). */
493 unsigned int has_backref : 1;
494 unsigned int has_constraint : 1;
495};
496typedef struct re_dfastate_t re_dfastate_t;
497
498struct re_state_table_entry
499{
500 int num;
501 int alloc;
502 re_dfastate_t **array;
503};
504
505/* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
506
507typedef struct
508{
509 int next_idx;
510 int alloc;
511 re_dfastate_t **array;
512} state_array_t;
513
514/* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
515
516typedef struct
517{
518 int node;
519 int str_idx; /* The position NODE match at. */
520 state_array_t path;
521} re_sub_match_last_t;
522
523/* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
524 And information about the node, whose type is OP_CLOSE_SUBEXP,
525 corresponding to NODE is stored in LASTS. */
526
527typedef struct
528{
529 int str_idx;
530 int node;
531 state_array_t *path;
532 int alasts; /* Allocation size of LASTS. */
533 int nlasts; /* The number of LASTS. */
534 re_sub_match_last_t **lasts;
535} re_sub_match_top_t;
536
537struct re_backref_cache_entry
538{
539 int node;
540 int str_idx;
541 int subexp_from;
542 int subexp_to;
543 char more;
544 char unused;
545 unsigned short int eps_reachable_subexps_map;
546};
547
548typedef struct
549{
550 /* The string object corresponding to the input string. */
551 re_string_t input;
552#if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
553 const re_dfa_t *const dfa;
554#else
555 const re_dfa_t *dfa;
556#endif
557 /* EFLAGS of the argument of regexec. */
558 int eflags;
559 /* Where the matching ends. */
560 int match_last;
561 int last_node;
562 /* The state log used by the matcher. */
563 re_dfastate_t **state_log;
564 int state_log_top;
565 /* Back reference cache. */
566 int nbkref_ents;
567 int abkref_ents;
568 struct re_backref_cache_entry *bkref_ents;
569 int max_mb_elem_len;
570 int nsub_tops;
571 int asub_tops;
572 re_sub_match_top_t **sub_tops;
573} re_match_context_t;
574
575typedef struct
576{
577 re_dfastate_t **sifted_states;
578 re_dfastate_t **limited_states;
579 int last_node;
580 int last_str_idx;
581 re_node_set limits;
582} re_sift_context_t;
583
584struct re_fail_stack_ent_t
585{
586 int idx;
587 int node;
588 regmatch_t *regs;
589 re_node_set eps_via_nodes;
590};
591
592struct re_fail_stack_t
593{
594 int num;
595 int alloc;
596 struct re_fail_stack_ent_t *stack;
597};
598
599struct re_dfa_t
600{
601 re_token_t *nodes;
602 size_t nodes_alloc;
603 size_t nodes_len;
604 int *nexts;
605 int *org_indices;
606 re_node_set *edests;
607 re_node_set *eclosures;
608 re_node_set *inveclosures;
609 struct re_state_table_entry *state_table;
610 re_dfastate_t *init_state;
611 re_dfastate_t *init_state_word;
612 re_dfastate_t *init_state_nl;
613 re_dfastate_t *init_state_begbuf;
614 bin_tree_t *str_tree;
615 bin_tree_storage_t *str_tree_storage;
616 re_bitset_ptr_t sb_char;
617 int str_tree_storage_idx;
618
619 /* number of subexpressions `re_nsub' is in regex_t. */
620 unsigned int state_hash_mask;
621 int init_node;
622 int nbackref; /* The number of backreference in this dfa. */
623
624 /* Bitmap expressing which backreference is used. */
625 bitset_word_t used_bkref_map;
626 bitset_word_t completed_bkref_map;
627
628 unsigned int has_plural_match : 1;
629 /* If this dfa has "multibyte node", which is a backreference or
630 a node which can accept multibyte character or multi character
631 collating element. */
632 unsigned int has_mb_node : 1;
633 unsigned int is_utf8 : 1;
634 unsigned int map_notascii : 1;
635 unsigned int word_ops_used : 1;
636 int mb_cur_max;
637 bitset_t word_char;
638 reg_syntax_t syntax;
639 int *subexp_map;
640#ifdef DEBUG
641 char* re_str;
642#endif
643 __libc_lock_define (, lock)
644};
645
646#define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
647#define re_node_set_remove(set,id) \
648 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
649#define re_node_set_empty(p) ((p)->nelem = 0)
650#define re_node_set_free(set) re_free ((set)->elems)
651
652
653typedef enum
654{
655 SB_CHAR,
656 MB_CHAR,
657 EQUIV_CLASS,
658 COLL_SYM,
659 CHAR_CLASS
660} bracket_elem_type;
661
662typedef struct
663{
664 bracket_elem_type type;
665 union
666 {
667 unsigned char ch;
668 unsigned char *name;
669 wchar_t wch;
670 } opr;
671} bracket_elem_t;
672
673
674/* Inline functions for bitset operation. */
675static void __attribute__ ((unused))
676bitset_not (bitset_t set)
677{
678 int bitset_i;
679 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
680 set[bitset_i] = ~set[bitset_i];
681}
682
683static void __attribute__ ((unused))
684bitset_merge (bitset_t dest, const bitset_t src)
685{
686 int bitset_i;
687 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
688 dest[bitset_i] |= src[bitset_i];
689}
690
691static void __attribute__ ((unused))
692bitset_mask (bitset_t dest, const bitset_t src)
693{
694 int bitset_i;
695 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
696 dest[bitset_i] &= src[bitset_i];
697}
698
699#ifdef RE_ENABLE_I18N
700/* Inline functions for re_string. */
701static int
702__attribute__ ((pure, unused))
703re_string_char_size_at (const re_string_t *pstr, int idx)
704{
705 int byte_idx;
706 if (pstr->mb_cur_max == 1)
707 return 1;
708 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
709 if (pstr->wcs[idx + byte_idx] != WEOF)
710 break;
711 return byte_idx;
712}
713
714static wint_t
715__attribute__ ((pure, unused))
716re_string_wchar_at (const re_string_t *pstr, int idx)
717{
718 if (pstr->mb_cur_max == 1)
719 return (wint_t) pstr->mbs[idx];
720 return (wint_t) pstr->wcs[idx];
721}
722
723# if IS_IN (libc)
724# ifdef _LIBC
725# include <locale/weight.h>
726# endif
727
728static int
729__attribute__ ((pure, unused))
730re_string_elem_size_at (const re_string_t *pstr, int idx)
731{
732# ifdef _LIBC
733 const unsigned char *p, *extra;
734 const int32_t *table, *indirect;
735 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
736
737 if (nrules != 0)
738 {
739 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
740 extra = (const unsigned char *)
741 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
742 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
743 _NL_COLLATE_INDIRECTMB);
744 p = pstr->mbs + idx;
745 findidx (table, indirect, extra, &p, pstr->len - idx);
746 return p - pstr->mbs - idx;
747 }
748 else
749# endif /* _LIBC */
750 return 1;
751}
752# endif
753#endif /* RE_ENABLE_I18N */
754
755#endif /* _REGEX_INTERNAL_H */
756