1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002, 2003, 2004, 2005 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
30 #if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET || defined _LIBC
31 # include <langinfo.h>
33 #if defined HAVE_LOCALE_H || defined _LIBC
36 #if defined HAVE_WCHAR_H || defined _LIBC
38 #endif /* HAVE_WCHAR_H || _LIBC */
39 #if defined HAVE_WCTYPE_H || defined _LIBC
41 #endif /* HAVE_WCTYPE_H || _LIBC */
43 # include <bits/libc-lock.h>
45 # define __libc_lock_define(CLASS,NAME)
46 # define __libc_lock_init(NAME) do { } while (0)
47 # define __libc_lock_lock(NAME) do { } while (0)
48 # define __libc_lock_unlock(NAME) do { } while (0)
51 /* In case that the system doesn't have isblank(). */
52 #if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank
53 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
57 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
58 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
59 # include <locale/localeinfo.h>
60 # include <locale/elem-hash.h>
61 # include <locale/coll-lookup.h>
65 /* This is for other GNU distributions with internationalized messages. */
66 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
70 # define gettext(msgid) \
71 INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES)
74 # define gettext(msgid) (msgid)
78 /* This define is so xgettext can find the internationalizable
80 # define gettext_noop(String) String
83 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC
84 # define RE_ENABLE_I18N
88 # define BE(expr, val) __builtin_expect (expr, val)
90 # define BE(expr, val) (expr)
94 /* Number of single byte character. */
97 #define COLL_ELEM_LEN_MAX 8
99 /* The character which represents newline. */
100 #define NEWLINE_CHAR '\n'
101 #define WIDE_NEWLINE_CHAR L'\n'
103 /* Rename to standard API for using out of glibc. */
105 # define __wctype wctype
106 # define __iswctype iswctype
107 # define __btowc btowc
109 # define __mempcpy mempcpy
111 # define __wcrtomb wcrtomb
112 # define __regfree regfree
113 # define attribute_hidden
114 #endif /* not _LIBC */
117 # define __attribute(arg) __attribute__ (arg)
119 # define __attribute(arg)
122 extern const char __re_error_msgid[] attribute_hidden;
123 extern const size_t __re_error_msgid_idx[] attribute_hidden;
125 typedef __re_idx_t Idx;
127 /* Special return value for failure to match. */
128 #define REG_MISSING ((Idx) -1)
130 /* Special return value for internal error. */
131 #define REG_ERROR ((Idx) -2)
133 /* Test whether N is a valid index, and is not one of the above. */
134 #ifdef _REGEX_LARGE_OFFSETS
135 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
137 # define REG_VALID_INDEX(n) (0 <= (n))
140 /* Test whether N is a valid nonzero index. */
141 #ifdef _REGEX_LARGE_OFFSETS
142 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
144 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
147 /* A hash value, suitable for computing hash tables. */
148 typedef __re_size_t re_hashval_t;
150 /* An integer used to represent a set of bits. It must be unsigned,
151 and must be at least as wide as unsigned int. */
152 typedef unsigned long int bitset_word;
154 /* Maximum value of a bitset word. It must be useful in preprocessor
155 contexts, and must be consistent with bitset_word. */
156 #define BITSET_WORD_MAX ULONG_MAX
158 /* Number of bits in a bitset word. Avoid greater-than-32-bit
159 integers and unconditional shifts by more than 31 bits, as they're
161 #if BITSET_WORD_MAX == 0xffffffff
162 # define BITSET_WORD_BITS 32
163 #elif BITSET_WORD_MAX >> 31 >> 5 == 1
164 # define BITSET_WORD_BITS 36
165 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
166 # define BITSET_WORD_BITS 48
167 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
168 # define BITSET_WORD_BITS 60
169 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
170 # define BITSET_WORD_BITS 64
171 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
172 # define BITSET_WORD_BITS 72
173 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
174 # define BITSET_WORD_BITS 128
175 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
176 # define BITSET_WORD_BITS 256
177 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
178 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
179 # if BITSET_WORD_BITS <= SBC_MAX
180 # error "Invalid SBC_MAX"
183 # error "Add case for new bitset_word size"
186 /* Number of bitset words in a bitset. */
187 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
189 typedef bitset_word bitset[BITSET_WORDS];
190 typedef bitset_word *re_bitset_ptr_t;
191 typedef const bitset_word *re_const_bitset_ptr_t;
193 #define PREV_WORD_CONSTRAINT 0x0001
194 #define PREV_NOTWORD_CONSTRAINT 0x0002
195 #define NEXT_WORD_CONSTRAINT 0x0004
196 #define NEXT_NOTWORD_CONSTRAINT 0x0008
197 #define PREV_NEWLINE_CONSTRAINT 0x0010
198 #define NEXT_NEWLINE_CONSTRAINT 0x0020
199 #define PREV_BEGBUF_CONSTRAINT 0x0040
200 #define NEXT_ENDBUF_CONSTRAINT 0x0080
201 #define WORD_DELIM_CONSTRAINT 0x0100
202 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
206 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
207 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
208 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
209 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
210 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
211 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
212 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
213 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
214 WORD_DELIM = WORD_DELIM_CONSTRAINT,
215 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
229 /* Node type, These are used by token, node, tree. */
235 #ifdef RE_ENABLE_I18N
238 #endif /* RE_ENABLE_I18N */
240 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
241 when the debugger shows values of this enum type. */
242 #define EPSILON_BIT 8
243 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
244 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
245 OP_ALT = EPSILON_BIT | 2,
246 OP_DUP_ASTERISK = EPSILON_BIT | 3,
247 ANCHOR = EPSILON_BIT | 4,
249 /* Tree type, these are used only by tree. */
253 /* Token type, these are used only by token. */
265 OP_CLOSE_EQUIV_CLASS,
276 #ifdef RE_ENABLE_I18N
279 /* Multibyte characters. */
282 /* Collating symbols. */
287 /* Equivalence classes. */
289 int32_t *equiv_classes;
292 /* Range expressions. */
294 uint32_t *range_starts;
295 uint32_t *range_ends;
296 # else /* not _LIBC */
297 wchar_t *range_starts;
299 # endif /* not _LIBC */
301 /* Character classes. */
302 wctype_t *char_classes;
304 /* If this character set is the non-matching list. */
305 unsigned int non_match : 1;
307 /* # of multibyte characters. */
310 /* # of collating symbols. */
313 /* # of equivalence classes. */
316 /* # of range expressions. */
319 /* # of character classes. */
322 #endif /* RE_ENABLE_I18N */
328 unsigned char c; /* for CHARACTER */
329 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
330 #ifdef RE_ENABLE_I18N
331 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
332 #endif /* RE_ENABLE_I18N */
333 Idx idx; /* for BACK_REF */
334 re_context_type ctx_type; /* for ANCHOR */
337 re_token_type_t type : 8;
339 re_token_type_t type;
341 unsigned int constraint : 10; /* context constraint */
342 unsigned int duplicated : 1;
343 unsigned int opt_subexp : 1;
344 #ifdef RE_ENABLE_I18N
345 unsigned int accept_mb : 1;
346 /* These 2 bits can be moved into the union if needed (e.g. if running out
347 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
348 unsigned int mb_partial : 1;
350 unsigned int word_char : 1;
353 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
357 /* Indicate the raw buffer which is the original string passed as an
358 argument of regexec(), re_search(), etc.. */
359 const unsigned char *raw_mbs;
360 /* Store the multibyte string. In case of "case insensitive mode" like
361 REG_ICASE, upper cases of the string are stored, otherwise MBS points
362 the same address that RAW_MBS points. */
364 #ifdef RE_ENABLE_I18N
365 /* Store the wide character string which is corresponding to MBS. */
370 /* Index in RAW_MBS. Each character mbs[i] corresponds to
371 raw_mbs[raw_mbs_idx + i]. */
373 /* The length of the valid characters in the buffers. */
375 /* The corresponding number of bytes in raw_mbs array. */
377 /* The length of the buffers MBS and WCS. */
379 /* The index in MBS, which is updated by re_string_fetch_byte. */
381 /* length of RAW_MBS array. */
383 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
385 /* End of the buffer may be shorter than its length in the cases such
386 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
389 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
392 /* The context of mbs[0]. We store the context independently, since
393 the context of mbs[0] may be different from raw_mbs[0], which is
394 the beginning of the input string. */
395 unsigned int tip_context;
396 /* The translation passed as a part of an argument of re_compile_pattern. */
397 unsigned REG_TRANSLATE_TYPE trans;
398 /* Copy of re_dfa_t's word_char. */
399 re_const_bitset_ptr_t word_char;
400 /* true if REG_ICASE. */
402 unsigned char is_utf8;
403 unsigned char map_notascii;
404 unsigned char mbs_allocated;
405 unsigned char offsets_needed;
406 unsigned char newline_anchor;
407 unsigned char word_ops_used;
410 typedef struct re_string_t re_string_t;
414 typedef struct re_dfa_t re_dfa_t;
418 # define internal_function __attribute ((regparm (3), stdcall))
420 # define internal_function
424 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
427 #ifdef RE_ENABLE_I18N
428 static void build_wcs_buffer (re_string_t *pstr) internal_function;
429 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
431 #endif /* RE_ENABLE_I18N */
432 static void build_upper_buffer (re_string_t *pstr) internal_function;
433 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
434 static unsigned int re_string_context_at (const re_string_t *input,
436 internal_function __attribute ((pure));
438 #define re_string_peek_byte(pstr, offset) \
439 ((pstr)->mbs[(pstr)->cur_idx + offset])
440 #define re_string_fetch_byte(pstr) \
441 ((pstr)->mbs[(pstr)->cur_idx++])
442 #define re_string_first_byte(pstr, idx) \
443 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
444 #define re_string_is_single_byte_char(pstr, idx) \
445 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
446 || (pstr)->wcs[(idx) + 1] != WEOF))
447 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
448 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
449 #define re_string_get_buffer(pstr) ((pstr)->mbs)
450 #define re_string_length(pstr) ((pstr)->len)
451 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
452 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
453 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
459 /* The OS usually guarantees only one guard page at the bottom of the stack,
460 and a page size can be as small as 4096 bytes. So we cannot safely
461 allocate anything larger than 4096 bytes. Also care for the possibility
462 of a few compiler-allocated temporary stack slots. */
463 # define __libc_use_alloca(n) ((n) < 4032)
465 /* alloca is implemented with malloc, so just use malloc. */
466 # define __libc_use_alloca(n) 0
470 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
471 #define re_xmalloc(t,n) ((t *) re_xnmalloc (n, sizeof (t)))
472 #define re_calloc(t,n) ((t *) calloc (n, sizeof (t)))
473 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
474 #define re_xrealloc(p,t,n) ((t *) re_xnrealloc (p, n, sizeof (t)))
475 #define re_x2realloc(p,t,pn) ((t *) re_x2nrealloc (p, pn, sizeof (t)))
476 #define re_free(p) free (p)
479 # define SIZE_MAX ((size_t) -1)
482 /* Return true if an array of N objects, each of size S, cannot exist
483 due to size arithmetic overflow. S must be nonzero. */
485 re_alloc_oversized (size_t n, size_t s)
487 return BE (SIZE_MAX / s < n, 0);
490 /* Return true if an array of (2 * N + 1) objects, each of size S,
491 cannot exist due to size arithmetic overflow. S must be nonzero. */
493 re_x2alloc_oversized (size_t n, size_t s)
495 return BE ((SIZE_MAX / s - 1) / 2 < n, 0);
498 /* Allocate an array of N objects, each with S bytes of memory,
499 dynamically, with error checking. S must be nonzero. */
501 re_xnmalloc (size_t n, size_t s)
503 return re_alloc_oversized (n, s) ? NULL : malloc (n * s);
506 /* Change the size of an allocated block of memory P to an array of N
507 objects each of S bytes, with error checking. S must be nonzero. */
509 re_xnrealloc (void *p, size_t n, size_t s)
511 return re_alloc_oversized (n, s) ? NULL : realloc (p, n * s);
514 /* Reallocate a block of memory P to an array of (2 * (*PN) + 1)
515 objects each of S bytes, with error checking. S must be nonzero.
516 If the allocation is successful, set *PN to the new allocation
517 count and return the resulting pointer. Otherwise, return
520 re_x2nrealloc (void *p, size_t *pn, size_t s)
522 if (re_x2alloc_oversized (*pn, s))
526 /* Add 1 in case *PN is zero. */
527 size_t n1 = 2 * *pn + 1;
528 p = realloc (p, n1 * s);
529 if (BE (p != NULL, 1))
537 struct bin_tree_t *parent;
538 struct bin_tree_t *left;
539 struct bin_tree_t *right;
540 struct bin_tree_t *first;
541 struct bin_tree_t *next;
545 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
546 Otherwise `type' indicate the type of this node. */
549 typedef struct bin_tree_t bin_tree_t;
551 #define BIN_TREE_STORAGE_SIZE \
552 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
554 struct bin_tree_storage_t
556 struct bin_tree_storage_t *next;
557 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
559 typedef struct bin_tree_storage_t bin_tree_storage_t;
561 #define CONTEXT_WORD 1
562 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
563 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
564 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
566 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
567 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
568 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
569 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
570 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
572 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
573 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
574 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
575 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
577 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
578 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
579 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
580 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
581 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
583 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
584 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
585 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
586 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
587 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
593 re_node_set non_eps_nodes;
594 re_node_set inveclosure;
595 re_node_set *entrance_nodes;
596 struct re_dfastate_t **trtable, **word_trtable;
597 unsigned int context : 4;
598 unsigned int halt : 1;
599 /* If this state can accept `multi byte'.
600 Note that we refer to multibyte characters, and multi character
601 collating elements as `multi byte'. */
602 unsigned int accept_mb : 1;
603 /* If this state has backreference node(s). */
604 unsigned int has_backref : 1;
605 unsigned int has_constraint : 1;
607 typedef struct re_dfastate_t re_dfastate_t;
609 struct re_state_table_entry
613 re_dfastate_t **array;
616 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
622 re_dfastate_t **array;
625 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
630 Idx str_idx; /* The position NODE match at. */
632 } re_sub_match_last_t;
634 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
635 And information about the node, whose type is OP_CLOSE_SUBEXP,
636 corresponding to NODE is stored in LASTS. */
643 Idx alasts; /* Allocation size of LASTS. */
644 Idx nlasts; /* The number of LASTS. */
645 re_sub_match_last_t **lasts;
646 } re_sub_match_top_t;
648 struct re_backref_cache_entry
656 unsigned short int eps_reachable_subexps_map;
661 /* The string object corresponding to the input string. */
663 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
668 /* EFLAGS of the argument of regexec. */
670 /* Where the matching ends. */
673 /* The state log used by the matcher. */
674 re_dfastate_t **state_log;
676 /* Back reference cache. */
679 struct re_backref_cache_entry *bkref_ents;
683 re_sub_match_top_t **sub_tops;
684 } re_match_context_t;
688 re_dfastate_t **sifted_states;
689 re_dfastate_t **limited_states;
695 struct re_fail_stack_ent_t
700 re_node_set eps_via_nodes;
703 struct re_fail_stack_t
707 struct re_fail_stack_ent_t *stack;
718 re_node_set *eclosures;
719 re_node_set *inveclosures;
720 struct re_state_table_entry *state_table;
721 re_dfastate_t *init_state;
722 re_dfastate_t *init_state_word;
723 re_dfastate_t *init_state_nl;
724 re_dfastate_t *init_state_begbuf;
725 bin_tree_t *str_tree;
726 bin_tree_storage_t *str_tree_storage;
727 re_bitset_ptr_t sb_char;
728 int str_tree_storage_idx;
730 /* number of subexpressions `re_nsub' is in regex_t. */
731 re_hashval_t state_hash_mask;
733 Idx nbackref; /* The number of backreference in this dfa. */
735 /* Bitmap expressing which backreference is used. */
736 bitset_word used_bkref_map;
737 bitset_word completed_bkref_map;
739 unsigned int has_plural_match : 1;
740 /* If this dfa has "multibyte node", which is a backreference or
741 a node which can accept multibyte character or multi character
742 collating element. */
743 unsigned int has_mb_node : 1;
744 unsigned int is_utf8 : 1;
745 unsigned int map_notascii : 1;
746 unsigned int word_ops_used : 1;
754 __libc_lock_define (, lock)
757 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
758 #define re_node_set_remove(set,id) \
759 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
760 #define re_node_set_empty(p) ((p)->nelem = 0)
761 #define re_node_set_free(set) re_free ((set)->elems)
763 static void free_state (re_dfastate_t *state) internal_function;
777 bracket_elem_type type;
787 /* Inline functions for bitset operation. */
790 bitset_set (bitset set, Idx i)
792 set[i / BITSET_WORD_BITS] |= (bitset_word) 1 << i % BITSET_WORD_BITS;
796 bitset_clear (bitset set, Idx i)
798 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word) 1 << i % BITSET_WORD_BITS);
802 bitset_contain (const bitset set, Idx i)
804 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
808 bitset_empty (bitset set)
810 memset (set, 0, sizeof (bitset));
814 bitset_set_all (bitset set)
816 memset (set, -1, sizeof (bitset_word) * (SBC_MAX / BITSET_WORD_BITS));
817 if (SBC_MAX % BITSET_WORD_BITS != 0)
818 set[BITSET_WORDS - 1] =
819 ((bitset_word) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
823 bitset_copy (bitset dest, const bitset src)
825 memcpy (dest, src, sizeof (bitset));
829 bitset_not (bitset set)
832 for (i = 0; i < SBC_MAX / BITSET_WORD_BITS; ++i)
834 if (SBC_MAX % BITSET_WORD_BITS != 0)
835 set[BITSET_WORDS - 1] =
836 (((bitset_word) 1 << SBC_MAX % BITSET_WORD_BITS) - 1
837 & ~set[BITSET_WORDS - 1]);
841 bitset_merge (bitset dest, const bitset src)
844 for (i = 0; i < BITSET_WORDS; ++i)
849 bitset_mask (bitset dest, const bitset src)
852 for (i = 0; i < BITSET_WORDS; ++i)
856 #if defined RE_ENABLE_I18N
857 /* Inline functions for re_string. */
859 internal_function __attribute ((pure))
860 re_string_char_size_at (const re_string_t *pstr, Idx idx)
863 if (pstr->mb_cur_max == 1)
865 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
866 if (pstr->wcs[idx + byte_idx] != WEOF)
872 internal_function __attribute ((pure))
873 re_string_wchar_at (const re_string_t *pstr, Idx idx)
875 if (pstr->mb_cur_max == 1)
876 return (wint_t) pstr->mbs[idx];
877 return (wint_t) pstr->wcs[idx];
881 internal_function __attribute ((pure))
882 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
885 const unsigned char *p, *extra;
886 const int32_t *table, *indirect;
888 # include <locale/weight.h>
889 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
893 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
894 extra = (const unsigned char *)
895 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
896 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
897 _NL_COLLATE_INDIRECTMB);
900 return p - pstr->mbs - idx;
906 #endif /* RE_ENABLE_I18N */
908 #endif /* _REGEX_INTERNAL_H */