1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002, 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4 Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
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)
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.
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. */
20 #ifndef _REGEX_INTERNAL_H
21 #define _REGEX_INTERNAL_H 1
34 #if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET || defined _LIBC
35 # include <langinfo.h>
37 #if defined HAVE_LOCALE_H || defined _LIBC
40 #if defined HAVE_WCHAR_H || defined _LIBC
42 #endif /* HAVE_WCHAR_H || _LIBC */
43 #if defined HAVE_WCTYPE_H || defined _LIBC
45 #endif /* HAVE_WCTYPE_H || _LIBC */
46 #if defined HAVE_STDINT_H || defined _LIBC
48 #endif /* HAVE_STDINT_H || _LIBC */
50 # include <bits/libc-lock.h>
52 # define __libc_lock_define(CLASS,NAME)
53 # define __libc_lock_init(NAME) do { } while (0)
54 # define __libc_lock_lock(NAME) do { } while (0)
55 # define __libc_lock_unlock(NAME) do { } while (0)
58 /* In case that the system doesn't have isblank(). */
59 #if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank
60 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
64 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
65 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
66 # include <locale/localeinfo.h>
67 # include <locale/elem-hash.h>
68 # include <locale/coll-lookup.h>
72 /* This is for other GNU distributions with internationalized messages. */
73 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
77 # define gettext(msgid) \
78 INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES)
81 # define gettext(msgid) (msgid)
85 /* This define is so xgettext can find the internationalizable
87 # define gettext_noop(String) String
90 /* For loser systems without the definition. */
92 # define SIZE_MAX ((size_t) -1)
95 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC
96 # define RE_ENABLE_I18N
100 # define BE(expr, val) __builtin_expect (expr, val)
102 # define BE(expr, val) (expr)
108 /* Number of ASCII characters. */
109 #define ASCII_CHARS 0x80
111 /* Number of single byte characters. */
112 #define SBC_MAX (UCHAR_MAX + 1)
114 #define COLL_ELEM_LEN_MAX 8
116 /* The character which represents newline. */
117 #define NEWLINE_CHAR '\n'
118 #define WIDE_NEWLINE_CHAR L'\n'
120 /* Rename to standard API for using out of glibc. */
122 # define __wctype wctype
123 # define __iswctype iswctype
124 # define __btowc btowc
126 # define __mempcpy mempcpy
128 # define __wcrtomb wcrtomb
129 # define __regfree regfree
130 # define attribute_hidden
131 #endif /* not _LIBC */
133 #if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
134 # define __attribute(arg) __attribute__ (arg)
136 # define __attribute(arg)
139 typedef __re_idx_t Idx;
141 /* Special return value for failure to match. */
142 #define REG_MISSING ((Idx) -1)
144 /* Special return value for internal error. */
145 #define REG_ERROR ((Idx) -2)
147 /* Test whether N is a valid index, and is not one of the above. */
148 #ifdef _REGEX_LARGE_OFFSETS
149 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
151 # define REG_VALID_INDEX(n) (0 <= (n))
154 /* Test whether N is a valid nonzero index. */
155 #ifdef _REGEX_LARGE_OFFSETS
156 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
158 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
161 /* A hash value, suitable for computing hash tables. */
162 typedef __re_size_t re_hashval_t;
164 /* An integer used to represent a set of bits. It must be unsigned,
165 and must be at least as wide as unsigned int. */
166 typedef unsigned long int bitset_word_t;
167 /* All bits set in a bitset_word_t. */
168 #define BITSET_WORD_MAX ULONG_MAX
170 /* Number of bits in a bitset_word_t. For portability to hosts with
171 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
172 instead, deduce it directly from BITSET_WORD_MAX. Avoid
173 greater-than-32-bit integers and unconditional shifts by more than
174 31 bits, as they're not portable. */
175 #if BITSET_WORD_MAX == 0xffffffff
176 # define BITSET_WORD_BITS 32
177 #elif BITSET_WORD_MAX >> 31 >> 5 == 1
178 # define BITSET_WORD_BITS 36
179 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
180 # define BITSET_WORD_BITS 48
181 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
182 # define BITSET_WORD_BITS 60
183 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
184 # define BITSET_WORD_BITS 64
185 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
186 # define BITSET_WORD_BITS 72
187 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
188 # define BITSET_WORD_BITS 128
189 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
190 # define BITSET_WORD_BITS 256
191 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
192 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
193 # if BITSET_WORD_BITS <= SBC_MAX
194 # error "Invalid SBC_MAX"
196 #elif BITSET_WORD_MAX == (0xffffffff + 2) * 0xffffffff
197 /* Work around a bug in 64-bit PGC (before version 6.1-2), where the
198 preprocessor mishandles large unsigned values as if they were signed. */
199 # define BITSET_WORD_BITS 64
201 # error "Add case for new bitset_word_t size"
204 /* Number of bitset_word_t values in a bitset_t. */
205 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
207 typedef bitset_word_t bitset_t[BITSET_WORDS];
208 typedef bitset_word_t *re_bitset_ptr_t;
209 typedef const bitset_word_t *re_const_bitset_ptr_t;
211 #define PREV_WORD_CONSTRAINT 0x0001
212 #define PREV_NOTWORD_CONSTRAINT 0x0002
213 #define NEXT_WORD_CONSTRAINT 0x0004
214 #define NEXT_NOTWORD_CONSTRAINT 0x0008
215 #define PREV_NEWLINE_CONSTRAINT 0x0010
216 #define NEXT_NEWLINE_CONSTRAINT 0x0020
217 #define PREV_BEGBUF_CONSTRAINT 0x0040
218 #define NEXT_ENDBUF_CONSTRAINT 0x0080
219 #define WORD_DELIM_CONSTRAINT 0x0100
220 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
224 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
225 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
226 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
227 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
228 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
229 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
230 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
231 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
232 WORD_DELIM = WORD_DELIM_CONSTRAINT,
233 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
247 /* Node type, These are used by token, node, tree. */
253 #ifdef RE_ENABLE_I18N
256 #endif /* RE_ENABLE_I18N */
258 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
259 when the debugger shows values of this enum type. */
260 #define EPSILON_BIT 8
261 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
262 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
263 OP_ALT = EPSILON_BIT | 2,
264 OP_DUP_ASTERISK = EPSILON_BIT | 3,
265 ANCHOR = EPSILON_BIT | 4,
267 /* Tree type, these are used only by tree. */
271 /* Token type, these are used only by token. */
283 OP_CLOSE_EQUIV_CLASS,
294 #ifdef RE_ENABLE_I18N
297 /* Multibyte characters. */
300 /* Collating symbols. */
305 /* Equivalence classes. */
307 int32_t *equiv_classes;
310 /* Range expressions. */
312 uint32_t *range_starts;
313 uint32_t *range_ends;
314 # else /* not _LIBC */
315 wchar_t *range_starts;
317 # endif /* not _LIBC */
319 /* Character classes. */
320 wctype_t *char_classes;
322 /* If this character set is the non-matching list. */
323 unsigned int non_match : 1;
325 /* # of multibyte characters. */
328 /* # of collating symbols. */
331 /* # of equivalence classes. */
334 /* # of range expressions. */
337 /* # of character classes. */
340 #endif /* RE_ENABLE_I18N */
346 unsigned char c; /* for CHARACTER */
347 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
348 #ifdef RE_ENABLE_I18N
349 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
350 #endif /* RE_ENABLE_I18N */
351 Idx idx; /* for BACK_REF */
352 re_context_type ctx_type; /* for ANCHOR */
355 re_token_type_t type : 8;
357 re_token_type_t type;
359 unsigned int constraint : 10; /* context constraint */
360 unsigned int duplicated : 1;
361 unsigned int opt_subexp : 1;
362 #ifdef RE_ENABLE_I18N
363 unsigned int accept_mb : 1;
364 /* These 2 bits can be moved into the union if needed (e.g. if running out
365 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
366 unsigned int mb_partial : 1;
368 unsigned int word_char : 1;
371 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
375 /* Indicate the raw buffer which is the original string passed as an
376 argument of regexec(), re_search(), etc.. */
377 const unsigned char *raw_mbs;
378 /* Store the multibyte string. In case of "case insensitive mode" like
379 REG_ICASE, upper cases of the string are stored, otherwise MBS points
380 the same address that RAW_MBS points. */
382 #ifdef RE_ENABLE_I18N
383 /* Store the wide character string which is corresponding to MBS. */
388 /* Index in RAW_MBS. Each character mbs[i] corresponds to
389 raw_mbs[raw_mbs_idx + i]. */
391 /* The length of the valid characters in the buffers. */
393 /* The corresponding number of bytes in raw_mbs array. */
395 /* The length of the buffers MBS and WCS. */
397 /* The index in MBS, which is updated by re_string_fetch_byte. */
399 /* length of RAW_MBS array. */
401 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
403 /* End of the buffer may be shorter than its length in the cases such
404 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
407 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
410 /* The context of mbs[0]. We store the context independently, since
411 the context of mbs[0] may be different from raw_mbs[0], which is
412 the beginning of the input string. */
413 unsigned int tip_context;
414 /* The translation passed as a part of an argument of re_compile_pattern. */
415 RE_TRANSLATE_TYPE trans;
416 /* Copy of re_dfa_t's word_char. */
417 re_const_bitset_ptr_t word_char;
418 /* true if REG_ICASE. */
420 unsigned char is_utf8;
421 unsigned char map_notascii;
422 unsigned char mbs_allocated;
423 unsigned char offsets_needed;
424 unsigned char newline_anchor;
425 unsigned char word_ops_used;
428 typedef struct re_string_t re_string_t;
432 typedef struct re_dfa_t re_dfa_t;
436 # define internal_function __attribute ((regparm (3), stdcall))
438 # define internal_function
442 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
445 #ifdef RE_ENABLE_I18N
446 static void build_wcs_buffer (re_string_t *pstr) internal_function;
447 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
449 #endif /* RE_ENABLE_I18N */
450 static void build_upper_buffer (re_string_t *pstr) internal_function;
451 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
452 static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
454 internal_function __attribute ((pure));
455 #define re_string_peek_byte(pstr, offset) \
456 ((pstr)->mbs[(pstr)->cur_idx + offset])
457 #define re_string_fetch_byte(pstr) \
458 ((pstr)->mbs[(pstr)->cur_idx++])
459 #define re_string_first_byte(pstr, idx) \
460 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
461 #define re_string_is_single_byte_char(pstr, idx) \
462 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
463 || (pstr)->wcs[(idx) + 1] != WEOF))
464 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
465 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
466 #define re_string_get_buffer(pstr) ((pstr)->mbs)
467 #define re_string_length(pstr) ((pstr)->len)
468 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
469 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
470 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
476 /* The OS usually guarantees only one guard page at the bottom of the stack,
477 and a page size can be as small as 4096 bytes. So we cannot safely
478 allocate anything larger than 4096 bytes. Also care for the possibility
479 of a few compiler-allocated temporary stack slots. */
480 # define __libc_use_alloca(n) ((n) < 4032)
482 /* alloca is implemented with malloc, so just use malloc. */
483 # define __libc_use_alloca(n) 0
488 # define MAX(a,b) ((a) < (b) ? (b) : (a))
491 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
492 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
493 #define re_free(p) free (p)
497 struct bin_tree_t *parent;
498 struct bin_tree_t *left;
499 struct bin_tree_t *right;
500 struct bin_tree_t *first;
501 struct bin_tree_t *next;
505 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
506 Otherwise `type' indicate the type of this node. */
509 typedef struct bin_tree_t bin_tree_t;
511 #define BIN_TREE_STORAGE_SIZE \
512 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
514 struct bin_tree_storage_t
516 struct bin_tree_storage_t *next;
517 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
519 typedef struct bin_tree_storage_t bin_tree_storage_t;
521 #define CONTEXT_WORD 1
522 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
523 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
524 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
526 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
527 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
528 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
529 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
530 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
532 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
533 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
534 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
535 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
537 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
538 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
539 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
540 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
541 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
543 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
544 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
545 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
546 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
547 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
553 re_node_set non_eps_nodes;
554 re_node_set inveclosure;
555 re_node_set *entrance_nodes;
556 struct re_dfastate_t **trtable, **word_trtable;
557 unsigned int context : 4;
558 unsigned int halt : 1;
559 /* If this state can accept `multi byte'.
560 Note that we refer to multibyte characters, and multi character
561 collating elements as `multi byte'. */
562 unsigned int accept_mb : 1;
563 /* If this state has backreference node(s). */
564 unsigned int has_backref : 1;
565 unsigned int has_constraint : 1;
567 typedef struct re_dfastate_t re_dfastate_t;
569 struct re_state_table_entry
573 re_dfastate_t **array;
576 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
582 re_dfastate_t **array;
585 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
590 Idx str_idx; /* The position NODE match at. */
592 } re_sub_match_last_t;
594 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
595 And information about the node, whose type is OP_CLOSE_SUBEXP,
596 corresponding to NODE is stored in LASTS. */
603 Idx alasts; /* Allocation size of LASTS. */
604 Idx nlasts; /* The number of LASTS. */
605 re_sub_match_last_t **lasts;
606 } re_sub_match_top_t;
608 struct re_backref_cache_entry
616 unsigned short int eps_reachable_subexps_map;
621 /* The string object corresponding to the input string. */
623 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
624 const re_dfa_t *const dfa;
628 /* EFLAGS of the argument of regexec. */
630 /* Where the matching ends. */
633 /* The state log used by the matcher. */
634 re_dfastate_t **state_log;
636 /* Back reference cache. */
639 struct re_backref_cache_entry *bkref_ents;
643 re_sub_match_top_t **sub_tops;
644 } re_match_context_t;
648 re_dfastate_t **sifted_states;
649 re_dfastate_t **limited_states;
655 struct re_fail_stack_ent_t
660 re_node_set eps_via_nodes;
663 struct re_fail_stack_t
667 struct re_fail_stack_ent_t *stack;
678 re_node_set *eclosures;
679 re_node_set *inveclosures;
680 struct re_state_table_entry *state_table;
681 re_dfastate_t *init_state;
682 re_dfastate_t *init_state_word;
683 re_dfastate_t *init_state_nl;
684 re_dfastate_t *init_state_begbuf;
685 bin_tree_t *str_tree;
686 bin_tree_storage_t *str_tree_storage;
687 re_bitset_ptr_t sb_char;
688 int str_tree_storage_idx;
690 /* number of subexpressions `re_nsub' is in regex_t. */
691 re_hashval_t state_hash_mask;
693 Idx nbackref; /* The number of backreference in this dfa. */
695 /* Bitmap expressing which backreference is used. */
696 bitset_word_t used_bkref_map;
697 bitset_word_t completed_bkref_map;
699 unsigned int has_plural_match : 1;
700 /* If this dfa has "multibyte node", which is a backreference or
701 a node which can accept multibyte character or multi character
702 collating element. */
703 unsigned int has_mb_node : 1;
704 unsigned int is_utf8 : 1;
705 unsigned int map_notascii : 1;
706 unsigned int word_ops_used : 1;
714 __libc_lock_define (, lock)
717 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
718 #define re_node_set_remove(set,id) \
719 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
720 #define re_node_set_empty(p) ((p)->nelem = 0)
721 #define re_node_set_free(set) re_free ((set)->elems)
735 bracket_elem_type type;
745 /* Inline functions for bitset_t operation. */
748 bitset_set (bitset_t set, Idx i)
750 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
754 bitset_clear (bitset_t set, Idx i)
756 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
760 bitset_contain (const bitset_t set, Idx i)
762 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
766 bitset_empty (bitset_t set)
768 memset (set, '\0', sizeof (bitset_t));
772 bitset_set_all (bitset_t set)
774 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
775 if (SBC_MAX % BITSET_WORD_BITS != 0)
776 set[BITSET_WORDS - 1] =
777 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
781 bitset_copy (bitset_t dest, const bitset_t src)
783 memcpy (dest, src, sizeof (bitset_t));
787 bitset_not (bitset_t set)
790 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
791 set[bitset_i] = ~set[bitset_i];
792 if (SBC_MAX % BITSET_WORD_BITS != 0)
793 set[BITSET_WORDS - 1] =
794 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
795 & ~set[BITSET_WORDS - 1]);
799 bitset_merge (bitset_t dest, const bitset_t src)
802 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
803 dest[bitset_i] |= src[bitset_i];
807 bitset_mask (bitset_t dest, const bitset_t src)
810 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
811 dest[bitset_i] &= src[bitset_i];
814 #ifdef RE_ENABLE_I18N
815 /* Inline functions for re_string. */
817 internal_function __attribute ((pure))
818 re_string_char_size_at (const re_string_t *pstr, Idx idx)
821 if (pstr->mb_cur_max == 1)
823 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
824 if (pstr->wcs[idx + byte_idx] != WEOF)
830 internal_function __attribute ((pure))
831 re_string_wchar_at (const re_string_t *pstr, Idx idx)
833 if (pstr->mb_cur_max == 1)
834 return (wint_t) pstr->mbs[idx];
835 return (wint_t) pstr->wcs[idx];
839 internal_function __attribute ((pure))
840 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
843 const unsigned char *p, *extra;
844 const int32_t *table, *indirect;
846 # include <locale/weight.h>
847 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
851 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
852 extra = (const unsigned char *)
853 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
854 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
855 _NL_COLLATE_INDIRECTMB);
858 return p - pstr->mbs - idx;
864 #endif /* RE_ENABLE_I18N */
866 #endif /* _REGEX_INTERNAL_H */