1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002-2012 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4 Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
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.
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.
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/>. */
20 #ifndef _REGEX_INTERNAL_H
21 #define _REGEX_INTERNAL_H 1
31 # include "localcharset.h"
39 # include <bits/libc-lock.h>
41 # define __libc_lock_define(CLASS,NAME)
42 # define __libc_lock_init(NAME) do { } while (0)
43 # define __libc_lock_lock(NAME) do { } while (0)
44 # define __libc_lock_unlock(NAME) do { } while (0)
47 /* In case that the system doesn't have isblank(). */
48 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
49 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
53 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
54 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
55 # include <locale/localeinfo.h>
56 # include <locale/elem-hash.h>
57 # include <locale/coll-lookup.h>
61 /* This is for other GNU distributions with internationalized messages. */
62 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
66 # define gettext(msgid) \
67 __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
70 # define gettext(msgid) (msgid)
74 /* This define is so xgettext can find the internationalizable
76 # define gettext_noop(String) String
79 #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE && HAVE_WCSCOLL) || _LIBC
80 # define RE_ENABLE_I18N
84 # define BE(expr, val) __builtin_expect (expr, val)
86 # define BE(expr, val) (expr)
89 /* Number of ASCII characters. */
90 #define ASCII_CHARS 0x80
92 /* Number of single byte characters. */
93 #define SBC_MAX (UCHAR_MAX + 1)
95 #define COLL_ELEM_LEN_MAX 8
97 /* The character which represents newline. */
98 #define NEWLINE_CHAR '\n'
99 #define WIDE_NEWLINE_CHAR L'\n'
101 /* Rename to standard API for using out of glibc. */
103 # define __wctype wctype
104 # define __iswctype iswctype
105 # define __btowc btowc
106 # define __mbrtowc mbrtowc
107 # define __wcrtomb wcrtomb
108 # define __regfree regfree
109 # define attribute_hidden
110 #endif /* not _LIBC */
112 #if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
113 # define __attribute(arg) __attribute__ (arg)
115 # define __attribute(arg)
118 typedef __re_idx_t Idx;
119 #ifdef _REGEX_LARGE_OFFSETS
120 # define IDX_MAX (SIZE_MAX - 2)
122 # define IDX_MAX INT_MAX
125 /* Special return value for failure to match. */
126 #define REG_MISSING ((Idx) -1)
128 /* Special return value for internal error. */
129 #define REG_ERROR ((Idx) -2)
131 /* Test whether N is a valid index, and is not one of the above. */
132 #ifdef _REGEX_LARGE_OFFSETS
133 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
135 # define REG_VALID_INDEX(n) (0 <= (n))
138 /* Test whether N is a valid nonzero index. */
139 #ifdef _REGEX_LARGE_OFFSETS
140 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
142 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
145 /* A hash value, suitable for computing hash tables. */
146 typedef __re_size_t re_hashval_t;
148 /* An integer used to represent a set of bits. It must be unsigned,
149 and must be at least as wide as unsigned int. */
150 typedef unsigned long int bitset_word_t;
151 /* All bits set in a bitset_word_t. */
152 #define BITSET_WORD_MAX ULONG_MAX
154 /* Number of bits in a bitset_word_t. For portability to hosts with
155 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
156 instead, deduce it directly from BITSET_WORD_MAX. Avoid
157 greater-than-32-bit integers and unconditional shifts by more than
158 31 bits, as they're not portable. */
159 #if BITSET_WORD_MAX == 0xffffffffUL
160 # define BITSET_WORD_BITS 32
161 #elif BITSET_WORD_MAX >> 31 >> 4 == 1
162 # define BITSET_WORD_BITS 36
163 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
164 # define BITSET_WORD_BITS 48
165 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
166 # define BITSET_WORD_BITS 60
167 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
168 # define BITSET_WORD_BITS 64
169 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
170 # define BITSET_WORD_BITS 72
171 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
172 # define BITSET_WORD_BITS 128
173 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
174 # define BITSET_WORD_BITS 256
175 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
176 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
177 # if BITSET_WORD_BITS <= SBC_MAX
178 # error "Invalid SBC_MAX"
181 # error "Add case for new bitset_word_t size"
184 /* Number of bitset_word_t values in a bitset_t. */
185 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
187 typedef bitset_word_t bitset_t[BITSET_WORDS];
188 typedef bitset_word_t *re_bitset_ptr_t;
189 typedef const bitset_word_t *re_const_bitset_ptr_t;
191 #define PREV_WORD_CONSTRAINT 0x0001
192 #define PREV_NOTWORD_CONSTRAINT 0x0002
193 #define NEXT_WORD_CONSTRAINT 0x0004
194 #define NEXT_NOTWORD_CONSTRAINT 0x0008
195 #define PREV_NEWLINE_CONSTRAINT 0x0010
196 #define NEXT_NEWLINE_CONSTRAINT 0x0020
197 #define PREV_BEGBUF_CONSTRAINT 0x0040
198 #define NEXT_ENDBUF_CONSTRAINT 0x0080
199 #define WORD_DELIM_CONSTRAINT 0x0100
200 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
204 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
205 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
206 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
207 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
208 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
209 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
210 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
211 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
212 WORD_DELIM = WORD_DELIM_CONSTRAINT,
213 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
227 /* Node type, These are used by token, node, tree. */
233 #ifdef RE_ENABLE_I18N
236 #endif /* RE_ENABLE_I18N */
238 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
239 when the debugger shows values of this enum type. */
240 #define EPSILON_BIT 8
241 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
242 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
243 OP_ALT = EPSILON_BIT | 2,
244 OP_DUP_ASTERISK = EPSILON_BIT | 3,
245 ANCHOR = EPSILON_BIT | 4,
247 /* Tree type, these are used only by tree. */
251 /* Token type, these are used only by token. */
263 OP_CLOSE_EQUIV_CLASS,
274 #ifdef RE_ENABLE_I18N
277 /* Multibyte characters. */
280 /* Collating symbols. */
285 /* Equivalence classes. */
287 int32_t *equiv_classes;
290 /* Range expressions. */
292 uint32_t *range_starts;
293 uint32_t *range_ends;
294 # else /* not _LIBC */
295 wchar_t *range_starts;
297 # endif /* not _LIBC */
299 /* Character classes. */
300 wctype_t *char_classes;
302 /* If this character set is the non-matching list. */
303 unsigned int non_match : 1;
305 /* # of multibyte characters. */
308 /* # of collating symbols. */
311 /* # of equivalence classes. */
314 /* # of range expressions. */
317 /* # of character classes. */
320 #endif /* RE_ENABLE_I18N */
326 unsigned char c; /* for CHARACTER */
327 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
328 #ifdef RE_ENABLE_I18N
329 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
330 #endif /* RE_ENABLE_I18N */
331 Idx idx; /* for BACK_REF */
332 re_context_type ctx_type; /* for ANCHOR */
334 #if __GNUC__ >= 2 && !defined __STRICT_ANSI__
335 re_token_type_t type : 8;
337 re_token_type_t type;
339 unsigned int constraint : 10; /* context constraint */
340 unsigned int duplicated : 1;
341 unsigned int opt_subexp : 1;
342 #ifdef RE_ENABLE_I18N
343 unsigned int accept_mb : 1;
344 /* These 2 bits can be moved into the union if needed (e.g. if running out
345 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
346 unsigned int mb_partial : 1;
348 unsigned int word_char : 1;
351 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
355 /* Indicate the raw buffer which is the original string passed as an
356 argument of regexec(), re_search(), etc.. */
357 const unsigned char *raw_mbs;
358 /* Store the multibyte string. In case of "case insensitive mode" like
359 REG_ICASE, upper cases of the string are stored, otherwise MBS points
360 the same address that RAW_MBS points. */
362 #ifdef RE_ENABLE_I18N
363 /* Store the wide character string which is corresponding to MBS. */
368 /* Index in RAW_MBS. Each character mbs[i] corresponds to
369 raw_mbs[raw_mbs_idx + i]. */
371 /* The length of the valid characters in the buffers. */
373 /* The corresponding number of bytes in raw_mbs array. */
375 /* The length of the buffers MBS and WCS. */
377 /* The index in MBS, which is updated by re_string_fetch_byte. */
379 /* length of RAW_MBS array. */
381 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
383 /* End of the buffer may be shorter than its length in the cases such
384 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
387 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
390 /* The context of mbs[0]. We store the context independently, since
391 the context of mbs[0] may be different from raw_mbs[0], which is
392 the beginning of the input string. */
393 unsigned int tip_context;
394 /* The translation passed as a part of an argument of re_compile_pattern. */
395 RE_TRANSLATE_TYPE trans;
396 /* Copy of re_dfa_t's word_char. */
397 re_const_bitset_ptr_t word_char;
398 /* true if REG_ICASE. */
400 unsigned char is_utf8;
401 unsigned char map_notascii;
402 unsigned char mbs_allocated;
403 unsigned char offsets_needed;
404 unsigned char newline_anchor;
405 unsigned char word_ops_used;
408 typedef struct re_string_t re_string_t;
412 typedef struct re_dfa_t re_dfa_t;
415 # define internal_function
419 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
422 # ifdef RE_ENABLE_I18N
423 static void build_wcs_buffer (re_string_t *pstr) internal_function;
424 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
426 # endif /* RE_ENABLE_I18N */
427 static void build_upper_buffer (re_string_t *pstr) internal_function;
428 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
429 static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
431 internal_function __attribute ((pure));
433 #define re_string_peek_byte(pstr, offset) \
434 ((pstr)->mbs[(pstr)->cur_idx + offset])
435 #define re_string_fetch_byte(pstr) \
436 ((pstr)->mbs[(pstr)->cur_idx++])
437 #define re_string_first_byte(pstr, idx) \
438 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
439 #define re_string_is_single_byte_char(pstr, idx) \
440 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
441 || (pstr)->wcs[(idx) + 1] != WEOF))
442 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
443 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
444 #define re_string_get_buffer(pstr) ((pstr)->mbs)
445 #define re_string_length(pstr) ((pstr)->len)
446 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
447 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
448 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
454 /* The OS usually guarantees only one guard page at the bottom of the stack,
455 and a page size can be as small as 4096 bytes. So we cannot safely
456 allocate anything larger than 4096 bytes. Also care for the possibility
457 of a few compiler-allocated temporary stack slots. */
458 # define __libc_use_alloca(n) ((n) < 4032)
460 /* alloca is implemented with malloc, so just use malloc. */
461 # define __libc_use_alloca(n) 0
463 # define alloca(n) malloc (n)
468 # define MALLOC_0_IS_NONNULL 1
469 #elif !defined MALLOC_0_IS_NONNULL
470 # define MALLOC_0_IS_NONNULL 0
474 # define MAX(a,b) ((a) < (b) ? (b) : (a))
477 # define MIN(a,b) ((a) < (b) ? (a) : (b))
480 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
481 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
482 #define re_free(p) free (p)
486 struct bin_tree_t *parent;
487 struct bin_tree_t *left;
488 struct bin_tree_t *right;
489 struct bin_tree_t *first;
490 struct bin_tree_t *next;
494 /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
495 Otherwise 'type' indicate the type of this node. */
498 typedef struct bin_tree_t bin_tree_t;
500 #define BIN_TREE_STORAGE_SIZE \
501 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
503 struct bin_tree_storage_t
505 struct bin_tree_storage_t *next;
506 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
508 typedef struct bin_tree_storage_t bin_tree_storage_t;
510 #define CONTEXT_WORD 1
511 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
512 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
513 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
515 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
516 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
517 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
518 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
519 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
521 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
522 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
523 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
524 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
526 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
527 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
528 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
529 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
530 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
532 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
533 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
534 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
535 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
536 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
542 re_node_set non_eps_nodes;
543 re_node_set inveclosure;
544 re_node_set *entrance_nodes;
545 struct re_dfastate_t **trtable, **word_trtable;
546 unsigned int context : 4;
547 unsigned int halt : 1;
548 /* If this state can accept "multi byte".
549 Note that we refer to multibyte characters, and multi character
550 collating elements as "multi byte". */
551 unsigned int accept_mb : 1;
552 /* If this state has backreference node(s). */
553 unsigned int has_backref : 1;
554 unsigned int has_constraint : 1;
556 typedef struct re_dfastate_t re_dfastate_t;
558 struct re_state_table_entry
562 re_dfastate_t **array;
565 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
571 re_dfastate_t **array;
574 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
579 Idx str_idx; /* The position NODE match at. */
581 } re_sub_match_last_t;
583 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
584 And information about the node, whose type is OP_CLOSE_SUBEXP,
585 corresponding to NODE is stored in LASTS. */
592 Idx alasts; /* Allocation size of LASTS. */
593 Idx nlasts; /* The number of LASTS. */
594 re_sub_match_last_t **lasts;
595 } re_sub_match_top_t;
597 struct re_backref_cache_entry
605 unsigned short int eps_reachable_subexps_map;
610 /* The string object corresponding to the input string. */
612 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
613 const re_dfa_t *const dfa;
617 /* EFLAGS of the argument of regexec. */
619 /* Where the matching ends. */
622 /* The state log used by the matcher. */
623 re_dfastate_t **state_log;
625 /* Back reference cache. */
628 struct re_backref_cache_entry *bkref_ents;
632 re_sub_match_top_t **sub_tops;
633 } re_match_context_t;
637 re_dfastate_t **sifted_states;
638 re_dfastate_t **limited_states;
644 struct re_fail_stack_ent_t
649 re_node_set eps_via_nodes;
652 struct re_fail_stack_t
656 struct re_fail_stack_ent_t *stack;
667 re_node_set *eclosures;
668 re_node_set *inveclosures;
669 struct re_state_table_entry *state_table;
670 re_dfastate_t *init_state;
671 re_dfastate_t *init_state_word;
672 re_dfastate_t *init_state_nl;
673 re_dfastate_t *init_state_begbuf;
674 bin_tree_t *str_tree;
675 bin_tree_storage_t *str_tree_storage;
676 re_bitset_ptr_t sb_char;
677 int str_tree_storage_idx;
679 /* number of subexpressions 're_nsub' is in regex_t. */
680 re_hashval_t state_hash_mask;
682 Idx nbackref; /* The number of backreference in this dfa. */
684 /* Bitmap expressing which backreference is used. */
685 bitset_word_t used_bkref_map;
686 bitset_word_t completed_bkref_map;
688 unsigned int has_plural_match : 1;
689 /* If this dfa has "multibyte node", which is a backreference or
690 a node which can accept multibyte character or multi character
691 collating element. */
692 unsigned int has_mb_node : 1;
693 unsigned int is_utf8 : 1;
694 unsigned int map_notascii : 1;
695 unsigned int word_ops_used : 1;
703 __libc_lock_define (, lock)
706 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
707 #define re_node_set_remove(set,id) \
708 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
709 #define re_node_set_empty(p) ((p)->nelem = 0)
710 #define re_node_set_free(set) re_free ((set)->elems)
724 bracket_elem_type type;
734 /* Functions for bitset_t operation. */
737 bitset_set (bitset_t set, Idx i)
739 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
743 bitset_clear (bitset_t set, Idx i)
745 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
749 bitset_contain (const bitset_t set, Idx i)
751 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
755 bitset_empty (bitset_t set)
757 memset (set, '\0', sizeof (bitset_t));
761 bitset_set_all (bitset_t set)
763 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
764 if (SBC_MAX % BITSET_WORD_BITS != 0)
765 set[BITSET_WORDS - 1] =
766 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
770 bitset_copy (bitset_t dest, const bitset_t src)
772 memcpy (dest, src, sizeof (bitset_t));
776 bitset_not (bitset_t set)
779 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
780 set[bitset_i] = ~set[bitset_i];
781 if (SBC_MAX % BITSET_WORD_BITS != 0)
782 set[BITSET_WORDS - 1] =
783 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
784 & ~set[BITSET_WORDS - 1]);
788 bitset_merge (bitset_t dest, const bitset_t src)
791 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
792 dest[bitset_i] |= src[bitset_i];
796 bitset_mask (bitset_t dest, const bitset_t src)
799 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
800 dest[bitset_i] &= src[bitset_i];
803 #ifdef RE_ENABLE_I18N
804 /* Functions for re_string. */
806 internal_function __attribute ((pure))
807 re_string_char_size_at (const re_string_t *pstr, Idx idx)
810 if (pstr->mb_cur_max == 1)
812 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
813 if (pstr->wcs[idx + byte_idx] != WEOF)
819 internal_function __attribute ((pure))
820 re_string_wchar_at (const re_string_t *pstr, Idx idx)
822 if (pstr->mb_cur_max == 1)
823 return (wint_t) pstr->mbs[idx];
824 return (wint_t) pstr->wcs[idx];
829 internal_function __attribute ((pure))
830 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
833 const unsigned char *p, *extra;
834 const int32_t *table, *indirect;
835 # include <locale/weight.h>
836 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
840 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
841 extra = (const unsigned char *)
842 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
843 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
844 _NL_COLLATE_INDIRECTMB);
846 findidx (&p, pstr->len - idx);
847 return p - pstr->mbs - idx;
854 #endif /* RE_ENABLE_I18N */
856 #ifndef __GNUC_PREREQ
857 # if defined __GNUC__ && defined __GNUC_MINOR__
858 # define __GNUC_PREREQ(maj, min) \
859 ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
861 # define __GNUC_PREREQ(maj, min) 0
865 #if __GNUC_PREREQ (3,4)
866 # undef __attribute_warn_unused_result__
867 # define __attribute_warn_unused_result__ \
868 __attribute__ ((__warn_unused_result__))
870 # define __attribute_warn_unused_result__ /* empty */
873 #endif /* _REGEX_INTERNAL_H */