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 Lesser General Public
8 License as published by the Free Software Foundation; either
9 version 2.1 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 Lesser General Public License for more details.
16 You should have received a copy of the GNU Lesser 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) 0
76 # define lock_lock(lock) ((void) 0)
77 # define lock_unlock(lock) ((void) 0)
80 /* In case that the system doesn't have isblank(). */
81 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
82 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
86 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
87 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
88 # include <locale/localeinfo.h>
89 # include <locale/elem-hash.h>
90 # include <locale/coll-lookup.h>
94 /* This is for other GNU distributions with internationalized messages. */
95 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
99 # define gettext(msgid) \
100 __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
103 # define gettext(msgid) (msgid)
107 /* This define is so xgettext can find the internationalizable
109 # define gettext_noop(String) String
112 #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE && HAVE_WCSCOLL) || _LIBC
113 # define RE_ENABLE_I18N
117 # define BE(expr, val) __builtin_expect (expr, val)
119 # define BE(expr, val) (expr)
122 /* Number of ASCII characters. */
123 #define ASCII_CHARS 0x80
125 /* Number of single byte characters. */
126 #define SBC_MAX (UCHAR_MAX + 1)
128 #define COLL_ELEM_LEN_MAX 8
130 /* The character which represents newline. */
131 #define NEWLINE_CHAR '\n'
132 #define WIDE_NEWLINE_CHAR L'\n'
134 /* Rename to standard API for using out of glibc. */
138 # define __wctype wctype
139 # define __iswctype iswctype
140 # define __btowc btowc
141 # define __mbrtowc mbrtowc
142 # define __wcrtomb wcrtomb
143 # define __regfree regfree
144 # define attribute_hidden
145 #endif /* not _LIBC */
147 #if __GNUC__ < 3 + (__GNUC_MINOR__ < 1)
148 # define __attribute__(arg)
151 typedef __re_idx_t Idx;
152 #ifdef _REGEX_LARGE_OFFSETS
153 # define IDX_MAX (SIZE_MAX - 2)
155 # define IDX_MAX INT_MAX
158 /* Special return value for failure to match. */
159 #define REG_MISSING ((Idx) -1)
161 /* Special return value for internal error. */
162 #define REG_ERROR ((Idx) -2)
164 /* Test whether N is a valid index, and is not one of the above. */
165 #ifdef _REGEX_LARGE_OFFSETS
166 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
168 # define REG_VALID_INDEX(n) (0 <= (n))
171 /* Test whether N is a valid nonzero index. */
172 #ifdef _REGEX_LARGE_OFFSETS
173 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
175 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
178 /* A hash value, suitable for computing hash tables. */
179 typedef __re_size_t re_hashval_t;
181 /* An integer used to represent a set of bits. It must be unsigned,
182 and must be at least as wide as unsigned int. */
183 typedef unsigned long int bitset_word_t;
184 /* All bits set in a bitset_word_t. */
185 #define BITSET_WORD_MAX ULONG_MAX
187 /* Number of bits in a bitset_word_t. For portability to hosts with
188 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
189 instead, deduce it directly from BITSET_WORD_MAX. Avoid
190 greater-than-32-bit integers and unconditional shifts by more than
191 31 bits, as they're not portable. */
192 #if BITSET_WORD_MAX == 0xffffffffUL
193 # define BITSET_WORD_BITS 32
194 #elif BITSET_WORD_MAX >> 31 >> 4 == 1
195 # define BITSET_WORD_BITS 36
196 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
197 # define BITSET_WORD_BITS 48
198 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
199 # define BITSET_WORD_BITS 60
200 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
201 # define BITSET_WORD_BITS 64
202 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
203 # define BITSET_WORD_BITS 72
204 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
205 # define BITSET_WORD_BITS 128
206 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
207 # define BITSET_WORD_BITS 256
208 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
209 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
210 # if BITSET_WORD_BITS <= SBC_MAX
211 # error "Invalid SBC_MAX"
214 # error "Add case for new bitset_word_t size"
217 /* Number of bitset_word_t values in a bitset_t. */
218 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
220 typedef bitset_word_t bitset_t[BITSET_WORDS];
221 typedef bitset_word_t *re_bitset_ptr_t;
222 typedef const bitset_word_t *re_const_bitset_ptr_t;
224 #define PREV_WORD_CONSTRAINT 0x0001
225 #define PREV_NOTWORD_CONSTRAINT 0x0002
226 #define NEXT_WORD_CONSTRAINT 0x0004
227 #define NEXT_NOTWORD_CONSTRAINT 0x0008
228 #define PREV_NEWLINE_CONSTRAINT 0x0010
229 #define NEXT_NEWLINE_CONSTRAINT 0x0020
230 #define PREV_BEGBUF_CONSTRAINT 0x0040
231 #define NEXT_ENDBUF_CONSTRAINT 0x0080
232 #define WORD_DELIM_CONSTRAINT 0x0100
233 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
237 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
238 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
239 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
240 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
241 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
242 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
243 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
244 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
245 WORD_DELIM = WORD_DELIM_CONSTRAINT,
246 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
260 /* Node type, These are used by token, node, tree. */
266 #ifdef RE_ENABLE_I18N
269 #endif /* RE_ENABLE_I18N */
271 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
272 when the debugger shows values of this enum type. */
273 #define EPSILON_BIT 8
274 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
275 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
276 OP_ALT = EPSILON_BIT | 2,
277 OP_DUP_ASTERISK = EPSILON_BIT | 3,
278 ANCHOR = EPSILON_BIT | 4,
280 /* Tree type, these are used only by tree. */
284 /* Token type, these are used only by token. */
296 OP_CLOSE_EQUIV_CLASS,
307 #ifdef RE_ENABLE_I18N
310 /* Multibyte characters. */
313 /* Collating symbols. */
318 /* Equivalence classes. */
320 int32_t *equiv_classes;
323 /* Range expressions. */
325 uint32_t *range_starts;
326 uint32_t *range_ends;
327 # else /* not _LIBC */
328 wchar_t *range_starts;
330 # endif /* not _LIBC */
332 /* Character classes. */
333 wctype_t *char_classes;
335 /* If this character set is the non-matching list. */
336 unsigned int non_match : 1;
338 /* # of multibyte characters. */
341 /* # of collating symbols. */
344 /* # of equivalence classes. */
347 /* # of range expressions. */
350 /* # of character classes. */
353 #endif /* RE_ENABLE_I18N */
359 unsigned char c; /* for CHARACTER */
360 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
361 #ifdef RE_ENABLE_I18N
362 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
363 #endif /* RE_ENABLE_I18N */
364 Idx idx; /* for BACK_REF */
365 re_context_type ctx_type; /* for ANCHOR */
367 #if __GNUC__ >= 2 && !defined __STRICT_ANSI__
368 re_token_type_t type : 8;
370 re_token_type_t type;
372 unsigned int constraint : 10; /* context constraint */
373 unsigned int duplicated : 1;
374 unsigned int opt_subexp : 1;
375 #ifdef RE_ENABLE_I18N
376 unsigned int accept_mb : 1;
377 /* These 2 bits can be moved into the union if needed (e.g. if running out
378 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
379 unsigned int mb_partial : 1;
381 unsigned int word_char : 1;
384 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
388 /* Indicate the raw buffer which is the original string passed as an
389 argument of regexec(), re_search(), etc.. */
390 const unsigned char *raw_mbs;
391 /* Store the multibyte string. In case of "case insensitive mode" like
392 REG_ICASE, upper cases of the string are stored, otherwise MBS points
393 the same address that RAW_MBS points. */
395 #ifdef RE_ENABLE_I18N
396 /* Store the wide character string which is corresponding to MBS. */
401 /* Index in RAW_MBS. Each character mbs[i] corresponds to
402 raw_mbs[raw_mbs_idx + i]. */
404 /* The length of the valid characters in the buffers. */
406 /* The corresponding number of bytes in raw_mbs array. */
408 /* The length of the buffers MBS and WCS. */
410 /* The index in MBS, which is updated by re_string_fetch_byte. */
412 /* length of RAW_MBS array. */
414 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
416 /* End of the buffer may be shorter than its length in the cases such
417 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
420 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
423 /* The context of mbs[0]. We store the context independently, since
424 the context of mbs[0] may be different from raw_mbs[0], which is
425 the beginning of the input string. */
426 unsigned int tip_context;
427 /* The translation passed as a part of an argument of re_compile_pattern. */
428 RE_TRANSLATE_TYPE trans;
429 /* Copy of re_dfa_t's word_char. */
430 re_const_bitset_ptr_t word_char;
431 /* true if REG_ICASE. */
433 unsigned char is_utf8;
434 unsigned char map_notascii;
435 unsigned char mbs_allocated;
436 unsigned char offsets_needed;
437 unsigned char newline_anchor;
438 unsigned char word_ops_used;
441 typedef struct re_string_t re_string_t;
445 typedef struct re_dfa_t re_dfa_t;
448 # define internal_function
452 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
455 # ifdef RE_ENABLE_I18N
456 static void build_wcs_buffer (re_string_t *pstr) internal_function;
457 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
459 # endif /* RE_ENABLE_I18N */
460 static void build_upper_buffer (re_string_t *pstr) internal_function;
461 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
462 static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
464 internal_function __attribute__ ((pure));
466 #define re_string_peek_byte(pstr, offset) \
467 ((pstr)->mbs[(pstr)->cur_idx + offset])
468 #define re_string_fetch_byte(pstr) \
469 ((pstr)->mbs[(pstr)->cur_idx++])
470 #define re_string_first_byte(pstr, idx) \
471 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
472 #define re_string_is_single_byte_char(pstr, idx) \
473 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
474 || (pstr)->wcs[(idx) + 1] != WEOF))
475 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
476 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
477 #define re_string_get_buffer(pstr) ((pstr)->mbs)
478 #define re_string_length(pstr) ((pstr)->len)
479 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
480 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
481 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
483 #if defined _LIBC || HAVE_ALLOCA
489 /* The OS usually guarantees only one guard page at the bottom of the stack,
490 and a page size can be as small as 4096 bytes. So we cannot safely
491 allocate anything larger than 4096 bytes. Also care for the possibility
492 of a few compiler-allocated temporary stack slots. */
493 # define __libc_use_alloca(n) ((n) < 4032)
495 /* alloca is implemented with malloc, so just use malloc. */
496 # define __libc_use_alloca(n) 0
498 # define alloca(n) malloc (n)
503 # define MALLOC_0_IS_NONNULL 1
504 #elif !defined MALLOC_0_IS_NONNULL
505 # define MALLOC_0_IS_NONNULL 0
509 # define MAX(a,b) ((a) < (b) ? (b) : (a))
512 # define MIN(a,b) ((a) < (b) ? (a) : (b))
515 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
516 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
517 #define re_free(p) free (p)
521 struct bin_tree_t *parent;
522 struct bin_tree_t *left;
523 struct bin_tree_t *right;
524 struct bin_tree_t *first;
525 struct bin_tree_t *next;
529 /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
530 Otherwise 'type' indicate the type of this node. */
533 typedef struct bin_tree_t bin_tree_t;
535 #define BIN_TREE_STORAGE_SIZE \
536 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
538 struct bin_tree_storage_t
540 struct bin_tree_storage_t *next;
541 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
543 typedef struct bin_tree_storage_t bin_tree_storage_t;
545 #define CONTEXT_WORD 1
546 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
547 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
548 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
550 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
551 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
552 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
553 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
554 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
556 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
557 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
558 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
559 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
561 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
562 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
563 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
564 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
565 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
567 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
568 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
569 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
570 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
571 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
577 re_node_set non_eps_nodes;
578 re_node_set inveclosure;
579 re_node_set *entrance_nodes;
580 struct re_dfastate_t **trtable, **word_trtable;
581 unsigned int context : 4;
582 unsigned int halt : 1;
583 /* If this state can accept "multi byte".
584 Note that we refer to multibyte characters, and multi character
585 collating elements as "multi byte". */
586 unsigned int accept_mb : 1;
587 /* If this state has backreference node(s). */
588 unsigned int has_backref : 1;
589 unsigned int has_constraint : 1;
591 typedef struct re_dfastate_t re_dfastate_t;
593 struct re_state_table_entry
597 re_dfastate_t **array;
600 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
606 re_dfastate_t **array;
609 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
614 Idx str_idx; /* The position NODE match at. */
616 } re_sub_match_last_t;
618 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
619 And information about the node, whose type is OP_CLOSE_SUBEXP,
620 corresponding to NODE is stored in LASTS. */
627 Idx alasts; /* Allocation size of LASTS. */
628 Idx nlasts; /* The number of LASTS. */
629 re_sub_match_last_t **lasts;
630 } re_sub_match_top_t;
632 struct re_backref_cache_entry
640 unsigned short int eps_reachable_subexps_map;
645 /* The string object corresponding to the input string. */
647 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
648 const re_dfa_t *const dfa;
652 /* EFLAGS of the argument of regexec. */
654 /* Where the matching ends. */
657 /* The state log used by the matcher. */
658 re_dfastate_t **state_log;
660 /* Back reference cache. */
663 struct re_backref_cache_entry *bkref_ents;
667 re_sub_match_top_t **sub_tops;
668 } re_match_context_t;
672 re_dfastate_t **sifted_states;
673 re_dfastate_t **limited_states;
679 struct re_fail_stack_ent_t
684 re_node_set eps_via_nodes;
687 struct re_fail_stack_t
691 struct re_fail_stack_ent_t *stack;
702 re_node_set *eclosures;
703 re_node_set *inveclosures;
704 struct re_state_table_entry *state_table;
705 re_dfastate_t *init_state;
706 re_dfastate_t *init_state_word;
707 re_dfastate_t *init_state_nl;
708 re_dfastate_t *init_state_begbuf;
709 bin_tree_t *str_tree;
710 bin_tree_storage_t *str_tree_storage;
711 re_bitset_ptr_t sb_char;
712 int str_tree_storage_idx;
714 /* number of subexpressions 're_nsub' is in regex_t. */
715 re_hashval_t state_hash_mask;
717 Idx nbackref; /* The number of backreference in this dfa. */
719 /* Bitmap expressing which backreference is used. */
720 bitset_word_t used_bkref_map;
721 bitset_word_t completed_bkref_map;
723 unsigned int has_plural_match : 1;
724 /* If this dfa has "multibyte node", which is a backreference or
725 a node which can accept multibyte character or multi character
726 collating element. */
727 unsigned int has_mb_node : 1;
728 unsigned int is_utf8 : 1;
729 unsigned int map_notascii : 1;
730 unsigned int word_ops_used : 1;
741 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
742 #define re_node_set_remove(set,id) \
743 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
744 #define re_node_set_empty(p) ((p)->nelem = 0)
745 #define re_node_set_free(set) re_free ((set)->elems)
759 bracket_elem_type type;
769 /* Functions for bitset_t operation. */
772 bitset_set (bitset_t set, Idx i)
774 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
778 bitset_clear (bitset_t set, Idx i)
780 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
784 bitset_contain (const bitset_t set, Idx i)
786 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
790 bitset_empty (bitset_t set)
792 memset (set, '\0', sizeof (bitset_t));
796 bitset_set_all (bitset_t set)
798 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
799 if (SBC_MAX % BITSET_WORD_BITS != 0)
800 set[BITSET_WORDS - 1] =
801 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
805 bitset_copy (bitset_t dest, const bitset_t src)
807 memcpy (dest, src, sizeof (bitset_t));
810 static void __attribute__ ((unused))
811 bitset_not (bitset_t set)
814 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
815 set[bitset_i] = ~set[bitset_i];
816 if (SBC_MAX % BITSET_WORD_BITS != 0)
817 set[BITSET_WORDS - 1] =
818 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
819 & ~set[BITSET_WORDS - 1]);
822 static void __attribute__ ((unused))
823 bitset_merge (bitset_t dest, const bitset_t src)
826 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
827 dest[bitset_i] |= src[bitset_i];
830 static void __attribute__ ((unused))
831 bitset_mask (bitset_t dest, const bitset_t src)
834 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
835 dest[bitset_i] &= src[bitset_i];
838 #ifdef RE_ENABLE_I18N
839 /* Functions for re_string. */
841 internal_function __attribute__ ((pure, unused))
842 re_string_char_size_at (const re_string_t *pstr, Idx idx)
845 if (pstr->mb_cur_max == 1)
847 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
848 if (pstr->wcs[idx + byte_idx] != WEOF)
854 internal_function __attribute__ ((pure, unused))
855 re_string_wchar_at (const re_string_t *pstr, Idx idx)
857 if (pstr->mb_cur_max == 1)
858 return (wint_t) pstr->mbs[idx];
859 return (wint_t) pstr->wcs[idx];
864 internal_function __attribute__ ((pure, unused))
865 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
868 const unsigned char *p, *extra;
869 const int32_t *table, *indirect;
870 # include <locale/weight.h>
871 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
875 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
876 extra = (const unsigned char *)
877 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
878 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
879 _NL_COLLATE_INDIRECTMB);
881 findidx (&p, pstr->len - idx);
882 return p - pstr->mbs - idx;
889 #endif /* RE_ENABLE_I18N */
891 #ifndef __GNUC_PREREQ
892 # if defined __GNUC__ && defined __GNUC_MINOR__
893 # define __GNUC_PREREQ(maj, min) \
894 ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
896 # define __GNUC_PREREQ(maj, min) 0
900 #if __GNUC_PREREQ (3,4)
901 # undef __attribute_warn_unused_result__
902 # define __attribute_warn_unused_result__ \
903 __attribute__ ((__warn_unused_result__))
905 # define __attribute_warn_unused_result__ /* empty */
908 #endif /* _REGEX_INTERNAL_H */