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 General Public
8 License as published by the Free Software Foundation; either
9 version 3 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 General Public License for more details.
16 You should have received a copy of the GNU 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
37 # include <bits/libc-lock.h>
38 # define lock_define(name) __libc_lock_define (, name)
39 # define lock_init(lock) (__libc_lock_init (lock), 0)
40 # define lock_fini(lock) 0
41 # define lock_lock(lock) __libc_lock_lock (lock)
42 # define lock_unlock(lock) __libc_lock_unlock (lock)
43 #elif defined GNULIB_LOCK
44 # include "glthread/lock.h"
45 /* Use gl_lock_define if empty macro arguments are known to work.
46 Otherwise, fall back on less-portable substitutes. */
47 # if ((defined __GNUC__ && !defined __STRICT_ANSI__) \
48 || (defined __STDC_VERSION__ && 199901L <= __STDC_VERSION__))
49 # define lock_define(name) gl_lock_define (, name)
50 # elif USE_POSIX_THREADS
51 # define lock_define(name) pthread_mutex_t name;
52 # elif USE_PTH_THREADS
53 # define lock_define(name) pth_mutex_t name;
54 # elif USE_SOLARIS_THREADS
55 # define lock_define(name) mutex_t name;
56 # elif USE_WINDOWS_THREADS
57 # define lock_define(name) gl_lock_t name;
59 # define lock_define(name)
61 # define lock_init(lock) glthread_lock_init (&(lock))
62 # define lock_fini(lock) glthread_lock_destroy (&(lock))
63 # define lock_lock(lock) glthread_lock_lock (&(lock))
64 # define lock_unlock(lock) glthread_lock_unlock (&(lock))
65 #elif defined GNULIB_PTHREAD
67 # define lock_define(name) pthread_mutex_t name;
68 # define lock_init(lock) pthread_mutex_init (&(lock), 0)
69 # define lock_fini(lock) pthread_mutex_destroy (&(lock))
70 # define lock_lock(lock) pthread_mutex_lock (&(lock))
71 # define lock_unlock(lock) pthread_mutex_unlock (&(lock))
73 # define lock_define(name)
74 # define lock_init(lock) 0
75 # define lock_fini(lock) ((void) 0)
76 /* The 'dfa' avoids an "unused variable 'dfa'" warning from GCC. */
77 # define lock_lock(lock) ((void) dfa)
78 # define lock_unlock(lock) ((void) 0)
81 /* In case that the system doesn't have isblank(). */
82 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
83 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
87 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
88 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
89 # include <locale/localeinfo.h>
90 # include <locale/elem-hash.h>
91 # include <locale/coll-lookup.h>
95 /* This is for other GNU distributions with internationalized messages. */
96 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
100 # define gettext(msgid) \
101 __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
104 # define gettext(msgid) (msgid)
108 /* This define is so xgettext can find the internationalizable
110 # define gettext_noop(String) String
113 #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE && HAVE_WCSCOLL) || _LIBC
114 # define RE_ENABLE_I18N
118 # define BE(expr, val) __builtin_expect (expr, val)
120 # define BE(expr, val) (expr)
123 /* Number of ASCII characters. */
124 #define ASCII_CHARS 0x80
126 /* Number of single byte characters. */
127 #define SBC_MAX (UCHAR_MAX + 1)
129 #define COLL_ELEM_LEN_MAX 8
131 /* The character which represents newline. */
132 #define NEWLINE_CHAR '\n'
133 #define WIDE_NEWLINE_CHAR L'\n'
135 /* Rename to standard API for using out of glibc. */
139 # define __wctype wctype
140 # define __iswctype iswctype
141 # define __btowc btowc
142 # define __mbrtowc mbrtowc
143 # define __wcrtomb wcrtomb
144 # define __regfree regfree
145 # define attribute_hidden
146 #endif /* not _LIBC */
148 #if __GNUC__ < 3 + (__GNUC_MINOR__ < 1)
149 # define __attribute__(arg)
152 typedef __re_idx_t Idx;
153 #ifdef _REGEX_LARGE_OFFSETS
154 # define IDX_MAX (SIZE_MAX - 2)
156 # define IDX_MAX INT_MAX
159 /* Special return value for failure to match. */
160 #define REG_MISSING ((Idx) -1)
162 /* Special return value for internal error. */
163 #define REG_ERROR ((Idx) -2)
165 /* Test whether N is a valid index, and is not one of the above. */
166 #ifdef _REGEX_LARGE_OFFSETS
167 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
169 # define REG_VALID_INDEX(n) (0 <= (n))
172 /* Test whether N is a valid nonzero index. */
173 #ifdef _REGEX_LARGE_OFFSETS
174 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
176 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
179 /* A hash value, suitable for computing hash tables. */
180 typedef __re_size_t re_hashval_t;
182 /* An integer used to represent a set of bits. It must be unsigned,
183 and must be at least as wide as unsigned int. */
184 typedef unsigned long int bitset_word_t;
185 /* All bits set in a bitset_word_t. */
186 #define BITSET_WORD_MAX ULONG_MAX
188 /* Number of bits in a bitset_word_t. For portability to hosts with
189 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
190 instead, deduce it directly from BITSET_WORD_MAX. Avoid
191 greater-than-32-bit integers and unconditional shifts by more than
192 31 bits, as they're not portable. */
193 #if BITSET_WORD_MAX == 0xffffffffUL
194 # define BITSET_WORD_BITS 32
195 #elif BITSET_WORD_MAX >> 31 >> 4 == 1
196 # define BITSET_WORD_BITS 36
197 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
198 # define BITSET_WORD_BITS 48
199 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
200 # define BITSET_WORD_BITS 60
201 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
202 # define BITSET_WORD_BITS 64
203 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
204 # define BITSET_WORD_BITS 72
205 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
206 # define BITSET_WORD_BITS 128
207 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
208 # define BITSET_WORD_BITS 256
209 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
210 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
211 # if BITSET_WORD_BITS <= SBC_MAX
212 # error "Invalid SBC_MAX"
215 # error "Add case for new bitset_word_t size"
218 /* Number of bitset_word_t values in a bitset_t. */
219 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
221 typedef bitset_word_t bitset_t[BITSET_WORDS];
222 typedef bitset_word_t *re_bitset_ptr_t;
223 typedef const bitset_word_t *re_const_bitset_ptr_t;
225 #define PREV_WORD_CONSTRAINT 0x0001
226 #define PREV_NOTWORD_CONSTRAINT 0x0002
227 #define NEXT_WORD_CONSTRAINT 0x0004
228 #define NEXT_NOTWORD_CONSTRAINT 0x0008
229 #define PREV_NEWLINE_CONSTRAINT 0x0010
230 #define NEXT_NEWLINE_CONSTRAINT 0x0020
231 #define PREV_BEGBUF_CONSTRAINT 0x0040
232 #define NEXT_ENDBUF_CONSTRAINT 0x0080
233 #define WORD_DELIM_CONSTRAINT 0x0100
234 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
238 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
239 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
240 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
241 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
242 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
243 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
244 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
245 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
246 WORD_DELIM = WORD_DELIM_CONSTRAINT,
247 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
261 /* Node type, These are used by token, node, tree. */
267 #ifdef RE_ENABLE_I18N
270 #endif /* RE_ENABLE_I18N */
272 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
273 when the debugger shows values of this enum type. */
274 #define EPSILON_BIT 8
275 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
276 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
277 OP_ALT = EPSILON_BIT | 2,
278 OP_DUP_ASTERISK = EPSILON_BIT | 3,
279 ANCHOR = EPSILON_BIT | 4,
281 /* Tree type, these are used only by tree. */
285 /* Token type, these are used only by token. */
297 OP_CLOSE_EQUIV_CLASS,
308 #ifdef RE_ENABLE_I18N
311 /* Multibyte characters. */
314 /* Collating symbols. */
319 /* Equivalence classes. */
321 int32_t *equiv_classes;
324 /* Range expressions. */
326 uint32_t *range_starts;
327 uint32_t *range_ends;
328 # else /* not _LIBC */
329 wchar_t *range_starts;
331 # endif /* not _LIBC */
333 /* Character classes. */
334 wctype_t *char_classes;
336 /* If this character set is the non-matching list. */
337 unsigned int non_match : 1;
339 /* # of multibyte characters. */
342 /* # of collating symbols. */
345 /* # of equivalence classes. */
348 /* # of range expressions. */
351 /* # of character classes. */
354 #endif /* RE_ENABLE_I18N */
360 unsigned char c; /* for CHARACTER */
361 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
362 #ifdef RE_ENABLE_I18N
363 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
364 #endif /* RE_ENABLE_I18N */
365 Idx idx; /* for BACK_REF */
366 re_context_type ctx_type; /* for ANCHOR */
368 #if __GNUC__ >= 2 && !defined __STRICT_ANSI__
369 re_token_type_t type : 8;
371 re_token_type_t type;
373 unsigned int constraint : 10; /* context constraint */
374 unsigned int duplicated : 1;
375 unsigned int opt_subexp : 1;
376 #ifdef RE_ENABLE_I18N
377 unsigned int accept_mb : 1;
378 /* These 2 bits can be moved into the union if needed (e.g. if running out
379 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
380 unsigned int mb_partial : 1;
382 unsigned int word_char : 1;
385 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
389 /* Indicate the raw buffer which is the original string passed as an
390 argument of regexec(), re_search(), etc.. */
391 const unsigned char *raw_mbs;
392 /* Store the multibyte string. In case of "case insensitive mode" like
393 REG_ICASE, upper cases of the string are stored, otherwise MBS points
394 the same address that RAW_MBS points. */
396 #ifdef RE_ENABLE_I18N
397 /* Store the wide character string which is corresponding to MBS. */
402 /* Index in RAW_MBS. Each character mbs[i] corresponds to
403 raw_mbs[raw_mbs_idx + i]. */
405 /* The length of the valid characters in the buffers. */
407 /* The corresponding number of bytes in raw_mbs array. */
409 /* The length of the buffers MBS and WCS. */
411 /* The index in MBS, which is updated by re_string_fetch_byte. */
413 /* length of RAW_MBS array. */
415 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
417 /* End of the buffer may be shorter than its length in the cases such
418 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
421 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
424 /* The context of mbs[0]. We store the context independently, since
425 the context of mbs[0] may be different from raw_mbs[0], which is
426 the beginning of the input string. */
427 unsigned int tip_context;
428 /* The translation passed as a part of an argument of re_compile_pattern. */
429 RE_TRANSLATE_TYPE trans;
430 /* Copy of re_dfa_t's word_char. */
431 re_const_bitset_ptr_t word_char;
432 /* true if REG_ICASE. */
434 unsigned char is_utf8;
435 unsigned char map_notascii;
436 unsigned char mbs_allocated;
437 unsigned char offsets_needed;
438 unsigned char newline_anchor;
439 unsigned char word_ops_used;
442 typedef struct re_string_t re_string_t;
446 typedef struct re_dfa_t re_dfa_t;
449 # define internal_function
453 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
456 # ifdef RE_ENABLE_I18N
457 static void build_wcs_buffer (re_string_t *pstr) internal_function;
458 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
460 # endif /* RE_ENABLE_I18N */
461 static void build_upper_buffer (re_string_t *pstr) internal_function;
462 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
463 static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
465 internal_function __attribute__ ((pure));
467 #define re_string_peek_byte(pstr, offset) \
468 ((pstr)->mbs[(pstr)->cur_idx + offset])
469 #define re_string_fetch_byte(pstr) \
470 ((pstr)->mbs[(pstr)->cur_idx++])
471 #define re_string_first_byte(pstr, idx) \
472 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
473 #define re_string_is_single_byte_char(pstr, idx) \
474 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
475 || (pstr)->wcs[(idx) + 1] != WEOF))
476 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
477 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
478 #define re_string_get_buffer(pstr) ((pstr)->mbs)
479 #define re_string_length(pstr) ((pstr)->len)
480 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
481 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
482 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
484 #if defined _LIBC || HAVE_ALLOCA
490 /* The OS usually guarantees only one guard page at the bottom of the stack,
491 and a page size can be as small as 4096 bytes. So we cannot safely
492 allocate anything larger than 4096 bytes. Also care for the possibility
493 of a few compiler-allocated temporary stack slots. */
494 # define __libc_use_alloca(n) ((n) < 4032)
496 /* alloca is implemented with malloc, so just use malloc. */
497 # define __libc_use_alloca(n) 0
499 # define alloca(n) malloc (n)
504 # define MALLOC_0_IS_NONNULL 1
505 #elif !defined MALLOC_0_IS_NONNULL
506 # define MALLOC_0_IS_NONNULL 0
510 # define MAX(a,b) ((a) < (b) ? (b) : (a))
513 # define MIN(a,b) ((a) < (b) ? (a) : (b))
516 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
517 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
518 #define re_free(p) free (p)
522 struct bin_tree_t *parent;
523 struct bin_tree_t *left;
524 struct bin_tree_t *right;
525 struct bin_tree_t *first;
526 struct bin_tree_t *next;
530 /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
531 Otherwise 'type' indicate the type of this node. */
534 typedef struct bin_tree_t bin_tree_t;
536 #define BIN_TREE_STORAGE_SIZE \
537 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
539 struct bin_tree_storage_t
541 struct bin_tree_storage_t *next;
542 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
544 typedef struct bin_tree_storage_t bin_tree_storage_t;
546 #define CONTEXT_WORD 1
547 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
548 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
549 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
551 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
552 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
553 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
554 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
555 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
557 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
558 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
559 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
560 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
562 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
563 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
564 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
565 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
566 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
568 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
569 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
570 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
571 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
572 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
578 re_node_set non_eps_nodes;
579 re_node_set inveclosure;
580 re_node_set *entrance_nodes;
581 struct re_dfastate_t **trtable, **word_trtable;
582 unsigned int context : 4;
583 unsigned int halt : 1;
584 /* If this state can accept "multi byte".
585 Note that we refer to multibyte characters, and multi character
586 collating elements as "multi byte". */
587 unsigned int accept_mb : 1;
588 /* If this state has backreference node(s). */
589 unsigned int has_backref : 1;
590 unsigned int has_constraint : 1;
592 typedef struct re_dfastate_t re_dfastate_t;
594 struct re_state_table_entry
598 re_dfastate_t **array;
601 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
607 re_dfastate_t **array;
610 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
615 Idx str_idx; /* The position NODE match at. */
617 } re_sub_match_last_t;
619 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
620 And information about the node, whose type is OP_CLOSE_SUBEXP,
621 corresponding to NODE is stored in LASTS. */
628 Idx alasts; /* Allocation size of LASTS. */
629 Idx nlasts; /* The number of LASTS. */
630 re_sub_match_last_t **lasts;
631 } re_sub_match_top_t;
633 struct re_backref_cache_entry
641 unsigned short int eps_reachable_subexps_map;
646 /* The string object corresponding to the input string. */
648 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
649 const re_dfa_t *const dfa;
653 /* EFLAGS of the argument of regexec. */
655 /* Where the matching ends. */
658 /* The state log used by the matcher. */
659 re_dfastate_t **state_log;
661 /* Back reference cache. */
664 struct re_backref_cache_entry *bkref_ents;
668 re_sub_match_top_t **sub_tops;
669 } re_match_context_t;
673 re_dfastate_t **sifted_states;
674 re_dfastate_t **limited_states;
680 struct re_fail_stack_ent_t
685 re_node_set eps_via_nodes;
688 struct re_fail_stack_t
692 struct re_fail_stack_ent_t *stack;
703 re_node_set *eclosures;
704 re_node_set *inveclosures;
705 struct re_state_table_entry *state_table;
706 re_dfastate_t *init_state;
707 re_dfastate_t *init_state_word;
708 re_dfastate_t *init_state_nl;
709 re_dfastate_t *init_state_begbuf;
710 bin_tree_t *str_tree;
711 bin_tree_storage_t *str_tree_storage;
712 re_bitset_ptr_t sb_char;
713 int str_tree_storage_idx;
715 /* number of subexpressions 're_nsub' is in regex_t. */
716 re_hashval_t state_hash_mask;
718 Idx nbackref; /* The number of backreference in this dfa. */
720 /* Bitmap expressing which backreference is used. */
721 bitset_word_t used_bkref_map;
722 bitset_word_t completed_bkref_map;
724 unsigned int has_plural_match : 1;
725 /* If this dfa has "multibyte node", which is a backreference or
726 a node which can accept multibyte character or multi character
727 collating element. */
728 unsigned int has_mb_node : 1;
729 unsigned int is_utf8 : 1;
730 unsigned int map_notascii : 1;
731 unsigned int word_ops_used : 1;
742 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
743 #define re_node_set_remove(set,id) \
744 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
745 #define re_node_set_empty(p) ((p)->nelem = 0)
746 #define re_node_set_free(set) re_free ((set)->elems)
760 bracket_elem_type type;
770 /* Functions for bitset_t operation. */
773 bitset_set (bitset_t set, Idx i)
775 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
779 bitset_clear (bitset_t set, Idx i)
781 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
785 bitset_contain (const bitset_t set, Idx i)
787 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
791 bitset_empty (bitset_t set)
793 memset (set, '\0', sizeof (bitset_t));
797 bitset_set_all (bitset_t set)
799 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
800 if (SBC_MAX % BITSET_WORD_BITS != 0)
801 set[BITSET_WORDS - 1] =
802 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
806 bitset_copy (bitset_t dest, const bitset_t src)
808 memcpy (dest, src, sizeof (bitset_t));
811 static void __attribute__ ((unused))
812 bitset_not (bitset_t set)
815 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
816 set[bitset_i] = ~set[bitset_i];
817 if (SBC_MAX % BITSET_WORD_BITS != 0)
818 set[BITSET_WORDS - 1] =
819 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
820 & ~set[BITSET_WORDS - 1]);
823 static void __attribute__ ((unused))
824 bitset_merge (bitset_t dest, const bitset_t src)
827 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
828 dest[bitset_i] |= src[bitset_i];
831 static void __attribute__ ((unused))
832 bitset_mask (bitset_t dest, const bitset_t src)
835 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
836 dest[bitset_i] &= src[bitset_i];
839 #ifdef RE_ENABLE_I18N
840 /* Functions for re_string. */
842 internal_function __attribute__ ((pure, unused))
843 re_string_char_size_at (const re_string_t *pstr, Idx idx)
846 if (pstr->mb_cur_max == 1)
848 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
849 if (pstr->wcs[idx + byte_idx] != WEOF)
855 internal_function __attribute__ ((pure, unused))
856 re_string_wchar_at (const re_string_t *pstr, Idx idx)
858 if (pstr->mb_cur_max == 1)
859 return (wint_t) pstr->mbs[idx];
860 return (wint_t) pstr->wcs[idx];
865 internal_function __attribute__ ((pure, unused))
866 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
869 const unsigned char *p, *extra;
870 const int32_t *table, *indirect;
871 # include <locale/weight.h>
872 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
876 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
877 extra = (const unsigned char *)
878 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
879 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
880 _NL_COLLATE_INDIRECTMB);
882 findidx (&p, pstr->len - idx);
883 return p - pstr->mbs - idx;
890 #endif /* RE_ENABLE_I18N */
892 #ifndef __GNUC_PREREQ
893 # if defined __GNUC__ && defined __GNUC_MINOR__
894 # define __GNUC_PREREQ(maj, min) \
895 ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
897 # define __GNUC_PREREQ(maj, min) 0
901 #if __GNUC_PREREQ (3,4)
902 # undef __attribute_warn_unused_result__
903 # define __attribute_warn_unused_result__ \
904 __attribute__ ((__warn_unused_result__))
906 # define __attribute_warn_unused_result__ /* empty */
909 #endif /* _REGEX_INTERNAL_H */