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_init(NAME) do { } while (0)
53 # define __libc_lock_lock(NAME) do { } while (0)
54 # define __libc_lock_unlock(NAME) do { } while (0)
57 /* In case that the system doesn't have isblank(). */
58 #if !defined _LIBC && !HAVE_DECL_ISBLANK && !defined isblank
59 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
63 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
64 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
65 # include <locale/localeinfo.h>
66 # include <locale/elem-hash.h>
67 # include <locale/coll-lookup.h>
71 /* This is for other GNU distributions with internationalized messages. */
72 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
76 # define gettext(msgid) \
77 INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES)
80 # define gettext(msgid) (msgid)
84 /* This define is so xgettext can find the internationalizable
86 # define gettext_noop(String) String
89 /* For loser systems without the definition. */
91 # define SIZE_MAX ((size_t) -1)
94 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC
95 # define RE_ENABLE_I18N
99 # define BE(expr, val) __builtin_expect (expr, val)
101 # define BE(expr, val) (expr)
107 /* Number of ASCII characters. */
108 #define ASCII_CHARS 0x80
110 /* Number of single byte characters. */
111 #define SBC_MAX (UCHAR_MAX + 1)
113 #define COLL_ELEM_LEN_MAX 8
115 /* The character which represents newline. */
116 #define NEWLINE_CHAR '\n'
117 #define WIDE_NEWLINE_CHAR L'\n'
119 /* Rename to standard API for using out of glibc. */
121 # define __wctype wctype
122 # define __iswctype iswctype
123 # define __btowc btowc
125 # define __mempcpy mempcpy
127 # define __wcrtomb wcrtomb
128 # define __regfree regfree
129 # define attribute_hidden
130 #endif /* not _LIBC */
132 #if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
133 # define __attribute(arg) __attribute__ (arg)
135 # define __attribute(arg)
138 typedef __re_idx_t Idx;
140 /* Special return value for failure to match. */
141 #define REG_MISSING ((Idx) -1)
143 /* Special return value for internal error. */
144 #define REG_ERROR ((Idx) -2)
146 /* Test whether N is a valid index, and is not one of the above. */
147 #ifdef _REGEX_LARGE_OFFSETS
148 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
150 # define REG_VALID_INDEX(n) (0 <= (n))
153 /* Test whether N is a valid nonzero index. */
154 #ifdef _REGEX_LARGE_OFFSETS
155 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
157 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
160 /* A hash value, suitable for computing hash tables. */
161 typedef __re_size_t re_hashval_t;
163 /* An integer used to represent a set of bits. It must be unsigned,
164 and must be at least as wide as unsigned int. */
165 typedef unsigned long int bitset_word_t;
166 /* All bits set in a bitset_word_t. */
167 #define BITSET_WORD_MAX ULONG_MAX
169 /* Number of bits in a bitset_word_t. For portability to hosts with
170 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
171 instead, deduce it directly from BITSET_WORD_MAX. Avoid
172 greater-than-32-bit integers and unconditional shifts by more than
173 31 bits, as they're not portable. */
174 #if BITSET_WORD_MAX == 0xffffffff
175 # define BITSET_WORD_BITS 32
176 #elif BITSET_WORD_MAX >> 31 >> 5 == 1
177 # define BITSET_WORD_BITS 36
178 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
179 # define BITSET_WORD_BITS 48
180 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
181 # define BITSET_WORD_BITS 60
182 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
183 # define BITSET_WORD_BITS 64
184 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
185 # define BITSET_WORD_BITS 72
186 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
187 # define BITSET_WORD_BITS 128
188 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
189 # define BITSET_WORD_BITS 256
190 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
191 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
192 # if BITSET_WORD_BITS <= SBC_MAX
193 # error "Invalid SBC_MAX"
195 #elif BITSET_WORD_MAX == (0xffffffff + 2) * 0xffffffff
196 /* Work around a bug in 64-bit PGC (before version 6.1-2), where the
197 preprocessor mishandles large unsigned values as if they were signed. */
198 # define BITSET_WORD_BITS 64
200 # error "Add case for new bitset_word_t size"
203 /* Number of bitset_word_t values in a bitset_t. */
204 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
206 typedef bitset_word_t bitset_t[BITSET_WORDS];
207 typedef bitset_word_t *re_bitset_ptr_t;
208 typedef const bitset_word_t *re_const_bitset_ptr_t;
210 #define PREV_WORD_CONSTRAINT 0x0001
211 #define PREV_NOTWORD_CONSTRAINT 0x0002
212 #define NEXT_WORD_CONSTRAINT 0x0004
213 #define NEXT_NOTWORD_CONSTRAINT 0x0008
214 #define PREV_NEWLINE_CONSTRAINT 0x0010
215 #define NEXT_NEWLINE_CONSTRAINT 0x0020
216 #define PREV_BEGBUF_CONSTRAINT 0x0040
217 #define NEXT_ENDBUF_CONSTRAINT 0x0080
218 #define WORD_DELIM_CONSTRAINT 0x0100
219 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
223 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
224 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
225 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
226 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
227 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
228 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
229 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
230 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
231 WORD_DELIM = WORD_DELIM_CONSTRAINT,
232 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
246 /* Node type, These are used by token, node, tree. */
252 #ifdef RE_ENABLE_I18N
255 #endif /* RE_ENABLE_I18N */
257 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
258 when the debugger shows values of this enum type. */
259 #define EPSILON_BIT 8
260 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
261 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
262 OP_ALT = EPSILON_BIT | 2,
263 OP_DUP_ASTERISK = EPSILON_BIT | 3,
264 ANCHOR = EPSILON_BIT | 4,
266 /* Tree type, these are used only by tree. */
270 /* Token type, these are used only by token. */
282 OP_CLOSE_EQUIV_CLASS,
293 #ifdef RE_ENABLE_I18N
296 /* Multibyte characters. */
299 /* Collating symbols. */
304 /* Equivalence classes. */
306 int32_t *equiv_classes;
309 /* Range expressions. */
311 uint32_t *range_starts;
312 uint32_t *range_ends;
313 # else /* not _LIBC */
314 wchar_t *range_starts;
316 # endif /* not _LIBC */
318 /* Character classes. */
319 wctype_t *char_classes;
321 /* If this character set is the non-matching list. */
322 unsigned int non_match : 1;
324 /* # of multibyte characters. */
327 /* # of collating symbols. */
330 /* # of equivalence classes. */
333 /* # of range expressions. */
336 /* # of character classes. */
339 #endif /* RE_ENABLE_I18N */
345 unsigned char c; /* for CHARACTER */
346 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
347 #ifdef RE_ENABLE_I18N
348 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
349 #endif /* RE_ENABLE_I18N */
350 Idx idx; /* for BACK_REF */
351 re_context_type ctx_type; /* for ANCHOR */
354 re_token_type_t type : 8;
356 re_token_type_t type;
358 unsigned int constraint : 10; /* context constraint */
359 unsigned int duplicated : 1;
360 unsigned int opt_subexp : 1;
361 #ifdef RE_ENABLE_I18N
362 unsigned int accept_mb : 1;
363 /* These 2 bits can be moved into the union if needed (e.g. if running out
364 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
365 unsigned int mb_partial : 1;
367 unsigned int word_char : 1;
370 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
374 /* Indicate the raw buffer which is the original string passed as an
375 argument of regexec(), re_search(), etc.. */
376 const unsigned char *raw_mbs;
377 /* Store the multibyte string. In case of "case insensitive mode" like
378 REG_ICASE, upper cases of the string are stored, otherwise MBS points
379 the same address that RAW_MBS points. */
381 #ifdef RE_ENABLE_I18N
382 /* Store the wide character string which is corresponding to MBS. */
387 /* Index in RAW_MBS. Each character mbs[i] corresponds to
388 raw_mbs[raw_mbs_idx + i]. */
390 /* The length of the valid characters in the buffers. */
392 /* The corresponding number of bytes in raw_mbs array. */
394 /* The length of the buffers MBS and WCS. */
396 /* The index in MBS, which is updated by re_string_fetch_byte. */
398 /* length of RAW_MBS array. */
400 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
402 /* End of the buffer may be shorter than its length in the cases such
403 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
406 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
409 /* The context of mbs[0]. We store the context independently, since
410 the context of mbs[0] may be different from raw_mbs[0], which is
411 the beginning of the input string. */
412 unsigned int tip_context;
413 /* The translation passed as a part of an argument of re_compile_pattern. */
414 RE_TRANSLATE_TYPE trans;
415 /* Copy of re_dfa_t's word_char. */
416 re_const_bitset_ptr_t word_char;
417 /* true if REG_ICASE. */
419 unsigned char is_utf8;
420 unsigned char map_notascii;
421 unsigned char mbs_allocated;
422 unsigned char offsets_needed;
423 unsigned char newline_anchor;
424 unsigned char word_ops_used;
427 typedef struct re_string_t re_string_t;
431 typedef struct re_dfa_t re_dfa_t;
435 # define internal_function __attribute ((regparm (3), stdcall))
437 # define internal_function
441 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
444 #ifdef RE_ENABLE_I18N
445 static void build_wcs_buffer (re_string_t *pstr) internal_function;
446 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
448 #endif /* RE_ENABLE_I18N */
449 static void build_upper_buffer (re_string_t *pstr) internal_function;
450 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
451 static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
453 internal_function __attribute ((pure));
454 #define re_string_peek_byte(pstr, offset) \
455 ((pstr)->mbs[(pstr)->cur_idx + offset])
456 #define re_string_fetch_byte(pstr) \
457 ((pstr)->mbs[(pstr)->cur_idx++])
458 #define re_string_first_byte(pstr, idx) \
459 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
460 #define re_string_is_single_byte_char(pstr, idx) \
461 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
462 || (pstr)->wcs[(idx) + 1] != WEOF))
463 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
464 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
465 #define re_string_get_buffer(pstr) ((pstr)->mbs)
466 #define re_string_length(pstr) ((pstr)->len)
467 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
468 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
469 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
475 /* The OS usually guarantees only one guard page at the bottom of the stack,
476 and a page size can be as small as 4096 bytes. So we cannot safely
477 allocate anything larger than 4096 bytes. Also care for the possibility
478 of a few compiler-allocated temporary stack slots. */
479 # define __libc_use_alloca(n) ((n) < 4032)
481 /* alloca is implemented with malloc, so just use malloc. */
482 # define __libc_use_alloca(n) 0
487 # define MAX(a,b) ((a) < (b) ? (b) : (a))
490 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
491 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
492 #define re_free(p) free (p)
496 struct bin_tree_t *parent;
497 struct bin_tree_t *left;
498 struct bin_tree_t *right;
499 struct bin_tree_t *first;
500 struct bin_tree_t *next;
504 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
505 Otherwise `type' indicate the type of this node. */
508 typedef struct bin_tree_t bin_tree_t;
510 #define BIN_TREE_STORAGE_SIZE \
511 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
513 struct bin_tree_storage_t
515 struct bin_tree_storage_t *next;
516 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
518 typedef struct bin_tree_storage_t bin_tree_storage_t;
520 #define CONTEXT_WORD 1
521 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
522 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
523 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
525 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
526 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
527 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
528 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
529 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
531 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
532 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
533 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
534 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
536 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
537 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
538 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
539 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
540 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
542 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
543 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
544 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
545 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
546 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
552 re_node_set non_eps_nodes;
553 re_node_set inveclosure;
554 re_node_set *entrance_nodes;
555 struct re_dfastate_t **trtable, **word_trtable;
556 unsigned int context : 4;
557 unsigned int halt : 1;
558 /* If this state can accept `multi byte'.
559 Note that we refer to multibyte characters, and multi character
560 collating elements as `multi byte'. */
561 unsigned int accept_mb : 1;
562 /* If this state has backreference node(s). */
563 unsigned int has_backref : 1;
564 unsigned int has_constraint : 1;
566 typedef struct re_dfastate_t re_dfastate_t;
568 struct re_state_table_entry
572 re_dfastate_t **array;
575 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
581 re_dfastate_t **array;
584 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
589 Idx str_idx; /* The position NODE match at. */
591 } re_sub_match_last_t;
593 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
594 And information about the node, whose type is OP_CLOSE_SUBEXP,
595 corresponding to NODE is stored in LASTS. */
602 Idx alasts; /* Allocation size of LASTS. */
603 Idx nlasts; /* The number of LASTS. */
604 re_sub_match_last_t **lasts;
605 } re_sub_match_top_t;
607 struct re_backref_cache_entry
615 unsigned short int eps_reachable_subexps_map;
620 /* The string object corresponding to the input string. */
622 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
623 const re_dfa_t *const dfa;
627 /* EFLAGS of the argument of regexec. */
629 /* Where the matching ends. */
632 /* The state log used by the matcher. */
633 re_dfastate_t **state_log;
635 /* Back reference cache. */
638 struct re_backref_cache_entry *bkref_ents;
642 re_sub_match_top_t **sub_tops;
643 } re_match_context_t;
647 re_dfastate_t **sifted_states;
648 re_dfastate_t **limited_states;
654 struct re_fail_stack_ent_t
659 re_node_set eps_via_nodes;
662 struct re_fail_stack_t
666 struct re_fail_stack_ent_t *stack;
677 re_node_set *eclosures;
678 re_node_set *inveclosures;
679 struct re_state_table_entry *state_table;
680 re_dfastate_t *init_state;
681 re_dfastate_t *init_state_word;
682 re_dfastate_t *init_state_nl;
683 re_dfastate_t *init_state_begbuf;
684 bin_tree_t *str_tree;
685 bin_tree_storage_t *str_tree_storage;
686 re_bitset_ptr_t sb_char;
687 int str_tree_storage_idx;
689 /* number of subexpressions `re_nsub' is in regex_t. */
690 re_hashval_t state_hash_mask;
692 Idx nbackref; /* The number of backreference in this dfa. */
694 /* Bitmap expressing which backreference is used. */
695 bitset_word_t used_bkref_map;
696 bitset_word_t completed_bkref_map;
698 unsigned int has_plural_match : 1;
699 /* If this dfa has "multibyte node", which is a backreference or
700 a node which can accept multibyte character or multi character
701 collating element. */
702 unsigned int has_mb_node : 1;
703 unsigned int is_utf8 : 1;
704 unsigned int map_notascii : 1;
705 unsigned int word_ops_used : 1;
714 __libc_lock_define (, lock)
718 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
719 #define re_node_set_remove(set,id) \
720 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
721 #define re_node_set_empty(p) ((p)->nelem = 0)
722 #define re_node_set_free(set) re_free ((set)->elems)
736 bracket_elem_type type;
746 /* Inline functions for bitset_t operation. */
749 bitset_set (bitset_t set, Idx i)
751 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
755 bitset_clear (bitset_t set, Idx i)
757 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
761 bitset_contain (const bitset_t set, Idx i)
763 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
767 bitset_empty (bitset_t set)
769 memset (set, '\0', sizeof (bitset_t));
773 bitset_set_all (bitset_t set)
775 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
776 if (SBC_MAX % BITSET_WORD_BITS != 0)
777 set[BITSET_WORDS - 1] =
778 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
782 bitset_copy (bitset_t dest, const bitset_t src)
784 memcpy (dest, src, sizeof (bitset_t));
788 bitset_not (bitset_t set)
791 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
792 set[bitset_i] = ~set[bitset_i];
793 if (SBC_MAX % BITSET_WORD_BITS != 0)
794 set[BITSET_WORDS - 1] =
795 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
796 & ~set[BITSET_WORDS - 1]);
800 bitset_merge (bitset_t dest, const bitset_t src)
803 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
804 dest[bitset_i] |= src[bitset_i];
808 bitset_mask (bitset_t dest, const bitset_t src)
811 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
812 dest[bitset_i] &= src[bitset_i];
815 #ifdef RE_ENABLE_I18N
816 /* Inline functions for re_string. */
818 internal_function __attribute ((pure))
819 re_string_char_size_at (const re_string_t *pstr, Idx idx)
822 if (pstr->mb_cur_max == 1)
824 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
825 if (pstr->wcs[idx + byte_idx] != WEOF)
831 internal_function __attribute ((pure))
832 re_string_wchar_at (const re_string_t *pstr, Idx idx)
834 if (pstr->mb_cur_max == 1)
835 return (wint_t) pstr->mbs[idx];
836 return (wint_t) pstr->wcs[idx];
840 internal_function __attribute ((pure))
841 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
844 const unsigned char *p, *extra;
845 const int32_t *table, *indirect;
847 # include <locale/weight.h>
848 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
852 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
853 extra = (const unsigned char *)
854 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
855 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
856 _NL_COLLATE_INDIRECTMB);
859 return p - pstr->mbs - idx;
865 #endif /* RE_ENABLE_I18N */
867 #endif /* _REGEX_INTERNAL_H */