1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002-2013 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
36 # include <bits/libc-lock.h>
38 # define __libc_lock_define(CLASS,NAME)
39 # define __libc_lock_init(NAME) do { } while (0)
40 # define __libc_lock_lock(NAME) do { } while (0)
41 # define __libc_lock_unlock(NAME) do { } while (0)
44 /* In case that the system doesn't have isblank(). */
45 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
46 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
50 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
51 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
52 # include <locale/localeinfo.h>
53 # include <locale/elem-hash.h>
54 # include <locale/coll-lookup.h>
58 /* This is for other GNU distributions with internationalized messages. */
59 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
63 # define gettext(msgid) \
64 __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
67 # define gettext(msgid) (msgid)
71 /* This define is so xgettext can find the internationalizable
73 # define gettext_noop(String) String
76 #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE && HAVE_WCSCOLL) || _LIBC
77 # define RE_ENABLE_I18N
81 # define BE(expr, val) __builtin_expect (expr, val)
83 # define BE(expr, val) (expr)
86 /* Number of ASCII characters. */
87 #define ASCII_CHARS 0x80
89 /* Number of single byte characters. */
90 #define SBC_MAX (UCHAR_MAX + 1)
92 #define COLL_ELEM_LEN_MAX 8
94 /* The character which represents newline. */
95 #define NEWLINE_CHAR '\n'
96 #define WIDE_NEWLINE_CHAR L'\n'
98 /* Rename to standard API for using out of glibc. */
102 # define __wctype wctype
103 # define __iswctype iswctype
104 # define __btowc btowc
105 # define __mbrtowc mbrtowc
106 # define __wcrtomb wcrtomb
107 # define __regfree regfree
108 # define attribute_hidden
109 #endif /* not _LIBC */
111 #if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
112 # define __attribute(arg) __attribute__ (arg)
114 # define __attribute(arg)
117 typedef __re_idx_t Idx;
118 #ifdef _REGEX_LARGE_OFFSETS
119 # define IDX_MAX (SIZE_MAX - 2)
121 # define IDX_MAX INT_MAX
124 /* Special return value for failure to match. */
125 #define REG_MISSING ((Idx) -1)
127 /* Special return value for internal error. */
128 #define REG_ERROR ((Idx) -2)
130 /* Test whether N is a valid index, and is not one of the above. */
131 #ifdef _REGEX_LARGE_OFFSETS
132 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
134 # define REG_VALID_INDEX(n) (0 <= (n))
137 /* Test whether N is a valid nonzero index. */
138 #ifdef _REGEX_LARGE_OFFSETS
139 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
141 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
144 /* A hash value, suitable for computing hash tables. */
145 typedef __re_size_t re_hashval_t;
147 /* An integer used to represent a set of bits. It must be unsigned,
148 and must be at least as wide as unsigned int. */
149 typedef unsigned long int bitset_word_t;
150 /* All bits set in a bitset_word_t. */
151 #define BITSET_WORD_MAX ULONG_MAX
153 /* Number of bits in a bitset_word_t. For portability to hosts with
154 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
155 instead, deduce it directly from BITSET_WORD_MAX. Avoid
156 greater-than-32-bit integers and unconditional shifts by more than
157 31 bits, as they're not portable. */
158 #if BITSET_WORD_MAX == 0xffffffffUL
159 # define BITSET_WORD_BITS 32
160 #elif BITSET_WORD_MAX >> 31 >> 4 == 1
161 # define BITSET_WORD_BITS 36
162 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
163 # define BITSET_WORD_BITS 48
164 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
165 # define BITSET_WORD_BITS 60
166 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
167 # define BITSET_WORD_BITS 64
168 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
169 # define BITSET_WORD_BITS 72
170 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
171 # define BITSET_WORD_BITS 128
172 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
173 # define BITSET_WORD_BITS 256
174 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
175 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
176 # if BITSET_WORD_BITS <= SBC_MAX
177 # error "Invalid SBC_MAX"
180 # error "Add case for new bitset_word_t size"
183 /* Number of bitset_word_t values in a bitset_t. */
184 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
186 typedef bitset_word_t bitset_t[BITSET_WORDS];
187 typedef bitset_word_t *re_bitset_ptr_t;
188 typedef const bitset_word_t *re_const_bitset_ptr_t;
190 #define PREV_WORD_CONSTRAINT 0x0001
191 #define PREV_NOTWORD_CONSTRAINT 0x0002
192 #define NEXT_WORD_CONSTRAINT 0x0004
193 #define NEXT_NOTWORD_CONSTRAINT 0x0008
194 #define PREV_NEWLINE_CONSTRAINT 0x0010
195 #define NEXT_NEWLINE_CONSTRAINT 0x0020
196 #define PREV_BEGBUF_CONSTRAINT 0x0040
197 #define NEXT_ENDBUF_CONSTRAINT 0x0080
198 #define WORD_DELIM_CONSTRAINT 0x0100
199 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
203 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
204 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
205 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
206 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
207 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
208 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
209 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
210 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
211 WORD_DELIM = WORD_DELIM_CONSTRAINT,
212 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
226 /* Node type, These are used by token, node, tree. */
232 #ifdef RE_ENABLE_I18N
235 #endif /* RE_ENABLE_I18N */
237 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
238 when the debugger shows values of this enum type. */
239 #define EPSILON_BIT 8
240 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
241 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
242 OP_ALT = EPSILON_BIT | 2,
243 OP_DUP_ASTERISK = EPSILON_BIT | 3,
244 ANCHOR = EPSILON_BIT | 4,
246 /* Tree type, these are used only by tree. */
250 /* Token type, these are used only by token. */
262 OP_CLOSE_EQUIV_CLASS,
273 #ifdef RE_ENABLE_I18N
276 /* Multibyte characters. */
279 /* Collating symbols. */
284 /* Equivalence classes. */
286 int32_t *equiv_classes;
289 /* Range expressions. */
291 uint32_t *range_starts;
292 uint32_t *range_ends;
293 # else /* not _LIBC */
294 wchar_t *range_starts;
296 # endif /* not _LIBC */
298 /* Character classes. */
299 wctype_t *char_classes;
301 /* If this character set is the non-matching list. */
302 unsigned int non_match : 1;
304 /* # of multibyte characters. */
307 /* # of collating symbols. */
310 /* # of equivalence classes. */
313 /* # of range expressions. */
316 /* # of character classes. */
319 #endif /* RE_ENABLE_I18N */
325 unsigned char c; /* for CHARACTER */
326 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
327 #ifdef RE_ENABLE_I18N
328 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
329 #endif /* RE_ENABLE_I18N */
330 Idx idx; /* for BACK_REF */
331 re_context_type ctx_type; /* for ANCHOR */
333 #if __GNUC__ >= 2 && !defined __STRICT_ANSI__
334 re_token_type_t type : 8;
336 re_token_type_t type;
338 unsigned int constraint : 10; /* context constraint */
339 unsigned int duplicated : 1;
340 unsigned int opt_subexp : 1;
341 #ifdef RE_ENABLE_I18N
342 unsigned int accept_mb : 1;
343 /* These 2 bits can be moved into the union if needed (e.g. if running out
344 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
345 unsigned int mb_partial : 1;
347 unsigned int word_char : 1;
350 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
354 /* Indicate the raw buffer which is the original string passed as an
355 argument of regexec(), re_search(), etc.. */
356 const unsigned char *raw_mbs;
357 /* Store the multibyte string. In case of "case insensitive mode" like
358 REG_ICASE, upper cases of the string are stored, otherwise MBS points
359 the same address that RAW_MBS points. */
361 #ifdef RE_ENABLE_I18N
362 /* Store the wide character string which is corresponding to MBS. */
367 /* Index in RAW_MBS. Each character mbs[i] corresponds to
368 raw_mbs[raw_mbs_idx + i]. */
370 /* The length of the valid characters in the buffers. */
372 /* The corresponding number of bytes in raw_mbs array. */
374 /* The length of the buffers MBS and WCS. */
376 /* The index in MBS, which is updated by re_string_fetch_byte. */
378 /* length of RAW_MBS array. */
380 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
382 /* End of the buffer may be shorter than its length in the cases such
383 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
386 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
389 /* The context of mbs[0]. We store the context independently, since
390 the context of mbs[0] may be different from raw_mbs[0], which is
391 the beginning of the input string. */
392 unsigned int tip_context;
393 /* The translation passed as a part of an argument of re_compile_pattern. */
394 RE_TRANSLATE_TYPE trans;
395 /* Copy of re_dfa_t's word_char. */
396 re_const_bitset_ptr_t word_char;
397 /* true if REG_ICASE. */
399 unsigned char is_utf8;
400 unsigned char map_notascii;
401 unsigned char mbs_allocated;
402 unsigned char offsets_needed;
403 unsigned char newline_anchor;
404 unsigned char word_ops_used;
407 typedef struct re_string_t re_string_t;
411 typedef struct re_dfa_t re_dfa_t;
414 # define internal_function
418 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
421 # ifdef RE_ENABLE_I18N
422 static void build_wcs_buffer (re_string_t *pstr) internal_function;
423 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
425 # endif /* RE_ENABLE_I18N */
426 static void build_upper_buffer (re_string_t *pstr) internal_function;
427 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
428 static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
430 internal_function __attribute ((pure));
432 #define re_string_peek_byte(pstr, offset) \
433 ((pstr)->mbs[(pstr)->cur_idx + offset])
434 #define re_string_fetch_byte(pstr) \
435 ((pstr)->mbs[(pstr)->cur_idx++])
436 #define re_string_first_byte(pstr, idx) \
437 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
438 #define re_string_is_single_byte_char(pstr, idx) \
439 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
440 || (pstr)->wcs[(idx) + 1] != WEOF))
441 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
442 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
443 #define re_string_get_buffer(pstr) ((pstr)->mbs)
444 #define re_string_length(pstr) ((pstr)->len)
445 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
446 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
447 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
453 /* The OS usually guarantees only one guard page at the bottom of the stack,
454 and a page size can be as small as 4096 bytes. So we cannot safely
455 allocate anything larger than 4096 bytes. Also care for the possibility
456 of a few compiler-allocated temporary stack slots. */
457 # define __libc_use_alloca(n) ((n) < 4032)
459 /* alloca is implemented with malloc, so just use malloc. */
460 # define __libc_use_alloca(n) 0
462 # define alloca(n) malloc (n)
467 # define MALLOC_0_IS_NONNULL 1
468 #elif !defined MALLOC_0_IS_NONNULL
469 # define MALLOC_0_IS_NONNULL 0
473 # define MAX(a,b) ((a) < (b) ? (b) : (a))
476 # define MIN(a,b) ((a) < (b) ? (a) : (b))
479 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
480 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
481 #define re_free(p) free (p)
485 struct bin_tree_t *parent;
486 struct bin_tree_t *left;
487 struct bin_tree_t *right;
488 struct bin_tree_t *first;
489 struct bin_tree_t *next;
493 /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
494 Otherwise 'type' indicate the type of this node. */
497 typedef struct bin_tree_t bin_tree_t;
499 #define BIN_TREE_STORAGE_SIZE \
500 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
502 struct bin_tree_storage_t
504 struct bin_tree_storage_t *next;
505 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
507 typedef struct bin_tree_storage_t bin_tree_storage_t;
509 #define CONTEXT_WORD 1
510 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
511 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
512 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
514 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
515 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
516 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
517 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
518 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
520 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
521 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
522 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
523 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
525 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
526 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
527 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
528 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
529 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
531 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
532 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
533 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
534 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
535 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
541 re_node_set non_eps_nodes;
542 re_node_set inveclosure;
543 re_node_set *entrance_nodes;
544 struct re_dfastate_t **trtable, **word_trtable;
545 unsigned int context : 4;
546 unsigned int halt : 1;
547 /* If this state can accept "multi byte".
548 Note that we refer to multibyte characters, and multi character
549 collating elements as "multi byte". */
550 unsigned int accept_mb : 1;
551 /* If this state has backreference node(s). */
552 unsigned int has_backref : 1;
553 unsigned int has_constraint : 1;
555 typedef struct re_dfastate_t re_dfastate_t;
557 struct re_state_table_entry
561 re_dfastate_t **array;
564 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
570 re_dfastate_t **array;
573 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
578 Idx str_idx; /* The position NODE match at. */
580 } re_sub_match_last_t;
582 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
583 And information about the node, whose type is OP_CLOSE_SUBEXP,
584 corresponding to NODE is stored in LASTS. */
591 Idx alasts; /* Allocation size of LASTS. */
592 Idx nlasts; /* The number of LASTS. */
593 re_sub_match_last_t **lasts;
594 } re_sub_match_top_t;
596 struct re_backref_cache_entry
604 unsigned short int eps_reachable_subexps_map;
609 /* The string object corresponding to the input string. */
611 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
612 const re_dfa_t *const dfa;
616 /* EFLAGS of the argument of regexec. */
618 /* Where the matching ends. */
621 /* The state log used by the matcher. */
622 re_dfastate_t **state_log;
624 /* Back reference cache. */
627 struct re_backref_cache_entry *bkref_ents;
631 re_sub_match_top_t **sub_tops;
632 } re_match_context_t;
636 re_dfastate_t **sifted_states;
637 re_dfastate_t **limited_states;
643 struct re_fail_stack_ent_t
648 re_node_set eps_via_nodes;
651 struct re_fail_stack_t
655 struct re_fail_stack_ent_t *stack;
666 re_node_set *eclosures;
667 re_node_set *inveclosures;
668 struct re_state_table_entry *state_table;
669 re_dfastate_t *init_state;
670 re_dfastate_t *init_state_word;
671 re_dfastate_t *init_state_nl;
672 re_dfastate_t *init_state_begbuf;
673 bin_tree_t *str_tree;
674 bin_tree_storage_t *str_tree_storage;
675 re_bitset_ptr_t sb_char;
676 int str_tree_storage_idx;
678 /* number of subexpressions 're_nsub' is in regex_t. */
679 re_hashval_t state_hash_mask;
681 Idx nbackref; /* The number of backreference in this dfa. */
683 /* Bitmap expressing which backreference is used. */
684 bitset_word_t used_bkref_map;
685 bitset_word_t completed_bkref_map;
687 unsigned int has_plural_match : 1;
688 /* If this dfa has "multibyte node", which is a backreference or
689 a node which can accept multibyte character or multi character
690 collating element. */
691 unsigned int has_mb_node : 1;
692 unsigned int is_utf8 : 1;
693 unsigned int map_notascii : 1;
694 unsigned int word_ops_used : 1;
702 __libc_lock_define (, lock)
705 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
706 #define re_node_set_remove(set,id) \
707 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
708 #define re_node_set_empty(p) ((p)->nelem = 0)
709 #define re_node_set_free(set) re_free ((set)->elems)
723 bracket_elem_type type;
733 /* Functions for bitset_t operation. */
736 bitset_set (bitset_t set, Idx i)
738 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
742 bitset_clear (bitset_t set, Idx i)
744 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
748 bitset_contain (const bitset_t set, Idx i)
750 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
754 bitset_empty (bitset_t set)
756 memset (set, '\0', sizeof (bitset_t));
760 bitset_set_all (bitset_t set)
762 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
763 if (SBC_MAX % BITSET_WORD_BITS != 0)
764 set[BITSET_WORDS - 1] =
765 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
769 bitset_copy (bitset_t dest, const bitset_t src)
771 memcpy (dest, src, sizeof (bitset_t));
775 bitset_not (bitset_t set)
778 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
779 set[bitset_i] = ~set[bitset_i];
780 if (SBC_MAX % BITSET_WORD_BITS != 0)
781 set[BITSET_WORDS - 1] =
782 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
783 & ~set[BITSET_WORDS - 1]);
787 bitset_merge (bitset_t dest, const bitset_t src)
790 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
791 dest[bitset_i] |= src[bitset_i];
795 bitset_mask (bitset_t dest, const bitset_t src)
798 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
799 dest[bitset_i] &= src[bitset_i];
802 #ifdef RE_ENABLE_I18N
803 /* Functions for re_string. */
805 internal_function __attribute ((pure))
806 re_string_char_size_at (const re_string_t *pstr, Idx idx)
809 if (pstr->mb_cur_max == 1)
811 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
812 if (pstr->wcs[idx + byte_idx] != WEOF)
818 internal_function __attribute ((pure))
819 re_string_wchar_at (const re_string_t *pstr, Idx idx)
821 if (pstr->mb_cur_max == 1)
822 return (wint_t) pstr->mbs[idx];
823 return (wint_t) pstr->wcs[idx];
828 internal_function __attribute ((pure))
829 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
832 const unsigned char *p, *extra;
833 const int32_t *table, *indirect;
834 # include <locale/weight.h>
835 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
839 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
840 extra = (const unsigned char *)
841 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
842 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
843 _NL_COLLATE_INDIRECTMB);
845 findidx (&p, pstr->len - idx);
846 return p - pstr->mbs - idx;
853 #endif /* RE_ENABLE_I18N */
855 #ifndef __GNUC_PREREQ
856 # if defined __GNUC__ && defined __GNUC_MINOR__
857 # define __GNUC_PREREQ(maj, min) \
858 ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
860 # define __GNUC_PREREQ(maj, min) 0
864 #if __GNUC_PREREQ (3,4)
865 # undef __attribute_warn_unused_result__
866 # define __attribute_warn_unused_result__ \
867 __attribute__ ((__warn_unused_result__))
869 # define __attribute_warn_unused_result__ /* empty */
872 #endif /* _REGEX_INTERNAL_H */