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 */
46 # include <bits/libc-lock.h>
48 # define __libc_lock_init(NAME) do { } while (0)
49 # define __libc_lock_lock(NAME) do { } while (0)
50 # define __libc_lock_unlock(NAME) do { } while (0)
53 /* In case that the system doesn't have isblank(). */
54 #if !defined _LIBC && !HAVE_DECL_ISBLANK && !defined isblank
55 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
59 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
60 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
61 # include <locale/localeinfo.h>
62 # include <locale/elem-hash.h>
63 # include <locale/coll-lookup.h>
67 /* This is for other GNU distributions with internationalized messages. */
68 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
72 # define gettext(msgid) \
73 INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES)
76 # define gettext(msgid) (msgid)
80 /* This define is so xgettext can find the internationalizable
82 # define gettext_noop(String) String
85 /* For loser systems without the definition. */
87 # define SIZE_MAX ((size_t) -1)
90 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_ISWCTYPE && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC
91 # define RE_ENABLE_I18N
95 # define BE(expr, val) __builtin_expect (expr, val)
97 # define BE(expr, val) (expr)
103 /* Number of ASCII characters. */
104 #define ASCII_CHARS 0x80
106 /* Number of single byte characters. */
107 #define SBC_MAX (UCHAR_MAX + 1)
109 #define COLL_ELEM_LEN_MAX 8
111 /* The character which represents newline. */
112 #define NEWLINE_CHAR '\n'
113 #define WIDE_NEWLINE_CHAR L'\n'
115 /* Rename to standard API for using out of glibc. */
117 # define __wctype wctype
118 # define __iswctype iswctype
119 # define __btowc btowc
121 # define __mempcpy mempcpy
123 # define __wcrtomb wcrtomb
124 # define __regfree regfree
125 # define attribute_hidden
126 #endif /* not _LIBC */
128 #if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
129 # define __attribute(arg) __attribute__ (arg)
131 # define __attribute(arg)
134 typedef __re_idx_t Idx;
136 /* Special return value for failure to match. */
137 #define REG_MISSING ((Idx) -1)
139 /* Special return value for internal error. */
140 #define REG_ERROR ((Idx) -2)
142 /* Test whether N is a valid index, and is not one of the above. */
143 #ifdef _REGEX_LARGE_OFFSETS
144 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
146 # define REG_VALID_INDEX(n) (0 <= (n))
149 /* Test whether N is a valid nonzero index. */
150 #ifdef _REGEX_LARGE_OFFSETS
151 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
153 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
156 /* A hash value, suitable for computing hash tables. */
157 typedef __re_size_t re_hashval_t;
159 /* An integer used to represent a set of bits. It must be unsigned,
160 and must be at least as wide as unsigned int. */
161 typedef unsigned long int bitset_word_t;
162 /* All bits set in a bitset_word_t. */
163 #define BITSET_WORD_MAX ULONG_MAX
165 /* Number of bits in a bitset_word_t. For portability to hosts with
166 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
167 instead, deduce it directly from BITSET_WORD_MAX. Avoid
168 greater-than-32-bit integers and unconditional shifts by more than
169 31 bits, as they're not portable. */
170 #if BITSET_WORD_MAX == 0xffffffff
171 # define BITSET_WORD_BITS 32
172 #elif BITSET_WORD_MAX >> 31 >> 5 == 1
173 # define BITSET_WORD_BITS 36
174 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
175 # define BITSET_WORD_BITS 48
176 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
177 # define BITSET_WORD_BITS 60
178 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
179 # define BITSET_WORD_BITS 64
180 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
181 # define BITSET_WORD_BITS 72
182 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
183 # define BITSET_WORD_BITS 128
184 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
185 # define BITSET_WORD_BITS 256
186 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
187 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
188 # if BITSET_WORD_BITS <= SBC_MAX
189 # error "Invalid SBC_MAX"
191 #elif BITSET_WORD_MAX == (0xffffffff + 2) * 0xffffffff
192 /* Work around a bug in 64-bit PGC (before version 6.1-2), where the
193 preprocessor mishandles large unsigned values as if they were signed. */
194 # define BITSET_WORD_BITS 64
196 # error "Add case for new bitset_word_t size"
199 /* Number of bitset_word_t values in a bitset_t. */
200 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
202 typedef bitset_word_t bitset_t[BITSET_WORDS];
203 typedef bitset_word_t *re_bitset_ptr_t;
204 typedef const bitset_word_t *re_const_bitset_ptr_t;
206 #define PREV_WORD_CONSTRAINT 0x0001
207 #define PREV_NOTWORD_CONSTRAINT 0x0002
208 #define NEXT_WORD_CONSTRAINT 0x0004
209 #define NEXT_NOTWORD_CONSTRAINT 0x0008
210 #define PREV_NEWLINE_CONSTRAINT 0x0010
211 #define NEXT_NEWLINE_CONSTRAINT 0x0020
212 #define PREV_BEGBUF_CONSTRAINT 0x0040
213 #define NEXT_ENDBUF_CONSTRAINT 0x0080
214 #define WORD_DELIM_CONSTRAINT 0x0100
215 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
219 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
220 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
221 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
222 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
223 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
224 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
225 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
226 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
227 WORD_DELIM = WORD_DELIM_CONSTRAINT,
228 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
242 /* Node type, These are used by token, node, tree. */
248 #ifdef RE_ENABLE_I18N
251 #endif /* RE_ENABLE_I18N */
253 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
254 when the debugger shows values of this enum type. */
255 #define EPSILON_BIT 8
256 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
257 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
258 OP_ALT = EPSILON_BIT | 2,
259 OP_DUP_ASTERISK = EPSILON_BIT | 3,
260 ANCHOR = EPSILON_BIT | 4,
262 /* Tree type, these are used only by tree. */
266 /* Token type, these are used only by token. */
278 OP_CLOSE_EQUIV_CLASS,
289 #ifdef RE_ENABLE_I18N
292 /* Multibyte characters. */
295 /* Collating symbols. */
300 /* Equivalence classes. */
302 int32_t *equiv_classes;
305 /* Range expressions. */
307 uint32_t *range_starts;
308 uint32_t *range_ends;
309 # else /* not _LIBC */
310 wchar_t *range_starts;
312 # endif /* not _LIBC */
314 /* Character classes. */
315 wctype_t *char_classes;
317 /* If this character set is the non-matching list. */
318 unsigned int non_match : 1;
320 /* # of multibyte characters. */
323 /* # of collating symbols. */
326 /* # of equivalence classes. */
329 /* # of range expressions. */
332 /* # of character classes. */
335 #endif /* RE_ENABLE_I18N */
341 unsigned char c; /* for CHARACTER */
342 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
343 #ifdef RE_ENABLE_I18N
344 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
345 #endif /* RE_ENABLE_I18N */
346 Idx idx; /* for BACK_REF */
347 re_context_type ctx_type; /* for ANCHOR */
349 #if __GNUC__ >= 2 && !__STRICT_ANSI__
350 re_token_type_t type : 8;
352 re_token_type_t type;
354 unsigned int constraint : 10; /* context constraint */
355 unsigned int duplicated : 1;
356 unsigned int opt_subexp : 1;
357 #ifdef RE_ENABLE_I18N
358 unsigned int accept_mb : 1;
359 /* These 2 bits can be moved into the union if needed (e.g. if running out
360 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
361 unsigned int mb_partial : 1;
363 unsigned int word_char : 1;
366 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
370 /* Indicate the raw buffer which is the original string passed as an
371 argument of regexec(), re_search(), etc.. */
372 const unsigned char *raw_mbs;
373 /* Store the multibyte string. In case of "case insensitive mode" like
374 REG_ICASE, upper cases of the string are stored, otherwise MBS points
375 the same address that RAW_MBS points. */
377 #ifdef RE_ENABLE_I18N
378 /* Store the wide character string which is corresponding to MBS. */
383 /* Index in RAW_MBS. Each character mbs[i] corresponds to
384 raw_mbs[raw_mbs_idx + i]. */
386 /* The length of the valid characters in the buffers. */
388 /* The corresponding number of bytes in raw_mbs array. */
390 /* The length of the buffers MBS and WCS. */
392 /* The index in MBS, which is updated by re_string_fetch_byte. */
394 /* length of RAW_MBS array. */
396 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
398 /* End of the buffer may be shorter than its length in the cases such
399 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
402 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
405 /* The context of mbs[0]. We store the context independently, since
406 the context of mbs[0] may be different from raw_mbs[0], which is
407 the beginning of the input string. */
408 unsigned int tip_context;
409 /* The translation passed as a part of an argument of re_compile_pattern. */
410 RE_TRANSLATE_TYPE trans;
411 /* Copy of re_dfa_t's word_char. */
412 re_const_bitset_ptr_t word_char;
413 /* true if REG_ICASE. */
415 unsigned char is_utf8;
416 unsigned char map_notascii;
417 unsigned char mbs_allocated;
418 unsigned char offsets_needed;
419 unsigned char newline_anchor;
420 unsigned char word_ops_used;
423 typedef struct re_string_t re_string_t;
427 typedef struct re_dfa_t re_dfa_t;
431 # define internal_function __attribute ((regparm (3), stdcall))
433 # define internal_function
437 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
440 #ifdef RE_ENABLE_I18N
441 static void build_wcs_buffer (re_string_t *pstr) internal_function;
442 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
444 #endif /* RE_ENABLE_I18N */
445 static void build_upper_buffer (re_string_t *pstr) internal_function;
446 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
447 static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
449 internal_function __attribute ((pure));
450 #define re_string_peek_byte(pstr, offset) \
451 ((pstr)->mbs[(pstr)->cur_idx + offset])
452 #define re_string_fetch_byte(pstr) \
453 ((pstr)->mbs[(pstr)->cur_idx++])
454 #define re_string_first_byte(pstr, idx) \
455 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
456 #define re_string_is_single_byte_char(pstr, idx) \
457 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
458 || (pstr)->wcs[(idx) + 1] != WEOF))
459 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
460 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
461 #define re_string_get_buffer(pstr) ((pstr)->mbs)
462 #define re_string_length(pstr) ((pstr)->len)
463 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
464 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
465 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
471 /* The OS usually guarantees only one guard page at the bottom of the stack,
472 and a page size can be as small as 4096 bytes. So we cannot safely
473 allocate anything larger than 4096 bytes. Also care for the possibility
474 of a few compiler-allocated temporary stack slots. */
475 # define __libc_use_alloca(n) ((n) < 4032)
477 /* alloca is implemented with malloc, so just use malloc. */
478 # define __libc_use_alloca(n) 0
483 # define MAX(a,b) ((a) < (b) ? (b) : (a))
486 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
487 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
488 #define re_free(p) free (p)
492 struct bin_tree_t *parent;
493 struct bin_tree_t *left;
494 struct bin_tree_t *right;
495 struct bin_tree_t *first;
496 struct bin_tree_t *next;
500 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
501 Otherwise `type' indicate the type of this node. */
504 typedef struct bin_tree_t bin_tree_t;
506 #define BIN_TREE_STORAGE_SIZE \
507 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
509 struct bin_tree_storage_t
511 struct bin_tree_storage_t *next;
512 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
514 typedef struct bin_tree_storage_t bin_tree_storage_t;
516 #define CONTEXT_WORD 1
517 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
518 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
519 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
521 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
522 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
523 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
524 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
525 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
527 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
528 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
529 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
530 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
532 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
533 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
534 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
535 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
536 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
538 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
539 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
540 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
541 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
542 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
548 re_node_set non_eps_nodes;
549 re_node_set inveclosure;
550 re_node_set *entrance_nodes;
551 struct re_dfastate_t **trtable, **word_trtable;
552 unsigned int context : 4;
553 unsigned int halt : 1;
554 /* If this state can accept `multi byte'.
555 Note that we refer to multibyte characters, and multi character
556 collating elements as `multi byte'. */
557 unsigned int accept_mb : 1;
558 /* If this state has backreference node(s). */
559 unsigned int has_backref : 1;
560 unsigned int has_constraint : 1;
562 typedef struct re_dfastate_t re_dfastate_t;
564 struct re_state_table_entry
568 re_dfastate_t **array;
571 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
577 re_dfastate_t **array;
580 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
585 Idx str_idx; /* The position NODE match at. */
587 } re_sub_match_last_t;
589 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
590 And information about the node, whose type is OP_CLOSE_SUBEXP,
591 corresponding to NODE is stored in LASTS. */
598 Idx alasts; /* Allocation size of LASTS. */
599 Idx nlasts; /* The number of LASTS. */
600 re_sub_match_last_t **lasts;
601 } re_sub_match_top_t;
603 struct re_backref_cache_entry
611 unsigned short int eps_reachable_subexps_map;
616 /* The string object corresponding to the input string. */
618 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
619 const re_dfa_t *const dfa;
623 /* EFLAGS of the argument of regexec. */
625 /* Where the matching ends. */
628 /* The state log used by the matcher. */
629 re_dfastate_t **state_log;
631 /* Back reference cache. */
634 struct re_backref_cache_entry *bkref_ents;
638 re_sub_match_top_t **sub_tops;
639 } re_match_context_t;
643 re_dfastate_t **sifted_states;
644 re_dfastate_t **limited_states;
650 struct re_fail_stack_ent_t
655 re_node_set eps_via_nodes;
658 struct re_fail_stack_t
662 struct re_fail_stack_ent_t *stack;
673 re_node_set *eclosures;
674 re_node_set *inveclosures;
675 struct re_state_table_entry *state_table;
676 re_dfastate_t *init_state;
677 re_dfastate_t *init_state_word;
678 re_dfastate_t *init_state_nl;
679 re_dfastate_t *init_state_begbuf;
680 bin_tree_t *str_tree;
681 bin_tree_storage_t *str_tree_storage;
682 re_bitset_ptr_t sb_char;
683 int str_tree_storage_idx;
685 /* number of subexpressions `re_nsub' is in regex_t. */
686 re_hashval_t state_hash_mask;
688 Idx nbackref; /* The number of backreference in this dfa. */
690 /* Bitmap expressing which backreference is used. */
691 bitset_word_t used_bkref_map;
692 bitset_word_t completed_bkref_map;
694 unsigned int has_plural_match : 1;
695 /* If this dfa has "multibyte node", which is a backreference or
696 a node which can accept multibyte character or multi character
697 collating element. */
698 unsigned int has_mb_node : 1;
699 unsigned int is_utf8 : 1;
700 unsigned int map_notascii : 1;
701 unsigned int word_ops_used : 1;
710 __libc_lock_define (, lock)
714 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
715 #define re_node_set_remove(set,id) \
716 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
717 #define re_node_set_empty(p) ((p)->nelem = 0)
718 #define re_node_set_free(set) re_free ((set)->elems)
732 bracket_elem_type type;
742 /* Inline functions for bitset_t operation. */
745 bitset_set (bitset_t set, Idx i)
747 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
751 bitset_clear (bitset_t set, Idx i)
753 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
757 bitset_contain (const bitset_t set, Idx i)
759 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
763 bitset_empty (bitset_t set)
765 memset (set, '\0', sizeof (bitset_t));
769 bitset_set_all (bitset_t set)
771 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
772 if (SBC_MAX % BITSET_WORD_BITS != 0)
773 set[BITSET_WORDS - 1] =
774 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
778 bitset_copy (bitset_t dest, const bitset_t src)
780 memcpy (dest, src, sizeof (bitset_t));
784 bitset_not (bitset_t set)
787 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
788 set[bitset_i] = ~set[bitset_i];
789 if (SBC_MAX % BITSET_WORD_BITS != 0)
790 set[BITSET_WORDS - 1] =
791 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
792 & ~set[BITSET_WORDS - 1]);
796 bitset_merge (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];
804 bitset_mask (bitset_t dest, const bitset_t src)
807 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
808 dest[bitset_i] &= src[bitset_i];
811 #ifdef RE_ENABLE_I18N
812 /* Inline functions for re_string. */
814 internal_function __attribute ((pure))
815 re_string_char_size_at (const re_string_t *pstr, Idx idx)
818 if (pstr->mb_cur_max == 1)
820 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
821 if (pstr->wcs[idx + byte_idx] != WEOF)
827 internal_function __attribute ((pure))
828 re_string_wchar_at (const re_string_t *pstr, Idx idx)
830 if (pstr->mb_cur_max == 1)
831 return (wint_t) pstr->mbs[idx];
832 return (wint_t) pstr->wcs[idx];
836 internal_function __attribute ((pure))
837 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
840 const unsigned char *p, *extra;
841 const int32_t *table, *indirect;
843 # include <locale/weight.h>
844 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
848 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
849 extra = (const unsigned char *)
850 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
851 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
852 _NL_COLLATE_INDIRECTMB);
855 return p - pstr->mbs - idx;
861 #endif /* RE_ENABLE_I18N */
863 #endif /* _REGEX_INTERNAL_H */