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>
39 /* Use __libc_lock_define (, NAME) if the library defines the macro,
40 and if the compiler is known to support empty macro arguments. */
41 #if (defined __libc_lock_define \
42 && ((defined __GNUC__ && !defined __STRICT_ANSI__) \
43 || (defined __STDC_VERSION__ && 199901L <= __STDC_VERSION__)))
44 # define lock_define(name) __libc_lock_define (, name)
45 # define lock_init(lock) (__libc_lock_init (lock), 0)
46 # define lock_fini(lock) 0
49 # define lock_define(name) pthread_mutex_t name;
50 # define lock_init(lock) pthread_mutex_init (&(lock), 0)
51 # define lock_fini(lock) pthread_mutex_destroy (&(lock))
52 # define __libc_lock_lock(lock) pthread_mutex_lock (&(lock))
53 # define __libc_lock_unlock(lock) pthread_mutex_unlock (&(lock))
56 /* In case that the system doesn't have isblank(). */
57 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
58 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
62 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
63 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
64 # include <locale/localeinfo.h>
65 # include <locale/elem-hash.h>
66 # include <locale/coll-lookup.h>
70 /* This is for other GNU distributions with internationalized messages. */
71 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
75 # define gettext(msgid) \
76 __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
79 # define gettext(msgid) (msgid)
83 /* This define is so xgettext can find the internationalizable
85 # define gettext_noop(String) String
88 #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE && HAVE_WCSCOLL) || _LIBC
89 # define RE_ENABLE_I18N
93 # define BE(expr, val) __builtin_expect (expr, val)
95 # define BE(expr, val) (expr)
98 /* Number of ASCII characters. */
99 #define ASCII_CHARS 0x80
101 /* Number of single byte characters. */
102 #define SBC_MAX (UCHAR_MAX + 1)
104 #define COLL_ELEM_LEN_MAX 8
106 /* The character which represents newline. */
107 #define NEWLINE_CHAR '\n'
108 #define WIDE_NEWLINE_CHAR L'\n'
110 /* Rename to standard API for using out of glibc. */
114 # define __wctype wctype
115 # define __iswctype iswctype
116 # define __btowc btowc
117 # define __mbrtowc mbrtowc
118 # define __wcrtomb wcrtomb
119 # define __regfree regfree
120 # define attribute_hidden
121 #endif /* not _LIBC */
123 #if __GNUC__ < 3 + (__GNUC_MINOR__ < 1)
124 # define __attribute__(arg)
127 typedef __re_idx_t Idx;
128 #ifdef _REGEX_LARGE_OFFSETS
129 # define IDX_MAX (SIZE_MAX - 2)
131 # define IDX_MAX INT_MAX
134 /* Special return value for failure to match. */
135 #define REG_MISSING ((Idx) -1)
137 /* Special return value for internal error. */
138 #define REG_ERROR ((Idx) -2)
140 /* Test whether N is a valid index, and is not one of the above. */
141 #ifdef _REGEX_LARGE_OFFSETS
142 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
144 # define REG_VALID_INDEX(n) (0 <= (n))
147 /* Test whether N is a valid nonzero index. */
148 #ifdef _REGEX_LARGE_OFFSETS
149 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
151 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
154 /* A hash value, suitable for computing hash tables. */
155 typedef __re_size_t re_hashval_t;
157 /* An integer used to represent a set of bits. It must be unsigned,
158 and must be at least as wide as unsigned int. */
159 typedef unsigned long int bitset_word_t;
160 /* All bits set in a bitset_word_t. */
161 #define BITSET_WORD_MAX ULONG_MAX
163 /* Number of bits in a bitset_word_t. For portability to hosts with
164 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
165 instead, deduce it directly from BITSET_WORD_MAX. Avoid
166 greater-than-32-bit integers and unconditional shifts by more than
167 31 bits, as they're not portable. */
168 #if BITSET_WORD_MAX == 0xffffffffUL
169 # define BITSET_WORD_BITS 32
170 #elif BITSET_WORD_MAX >> 31 >> 4 == 1
171 # define BITSET_WORD_BITS 36
172 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
173 # define BITSET_WORD_BITS 48
174 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
175 # define BITSET_WORD_BITS 60
176 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
177 # define BITSET_WORD_BITS 64
178 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
179 # define BITSET_WORD_BITS 72
180 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
181 # define BITSET_WORD_BITS 128
182 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
183 # define BITSET_WORD_BITS 256
184 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
185 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
186 # if BITSET_WORD_BITS <= SBC_MAX
187 # error "Invalid SBC_MAX"
190 # error "Add case for new bitset_word_t size"
193 /* Number of bitset_word_t values in a bitset_t. */
194 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
196 typedef bitset_word_t bitset_t[BITSET_WORDS];
197 typedef bitset_word_t *re_bitset_ptr_t;
198 typedef const bitset_word_t *re_const_bitset_ptr_t;
200 #define PREV_WORD_CONSTRAINT 0x0001
201 #define PREV_NOTWORD_CONSTRAINT 0x0002
202 #define NEXT_WORD_CONSTRAINT 0x0004
203 #define NEXT_NOTWORD_CONSTRAINT 0x0008
204 #define PREV_NEWLINE_CONSTRAINT 0x0010
205 #define NEXT_NEWLINE_CONSTRAINT 0x0020
206 #define PREV_BEGBUF_CONSTRAINT 0x0040
207 #define NEXT_ENDBUF_CONSTRAINT 0x0080
208 #define WORD_DELIM_CONSTRAINT 0x0100
209 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
213 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
214 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
215 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
216 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
217 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
218 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
219 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
220 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
221 WORD_DELIM = WORD_DELIM_CONSTRAINT,
222 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
236 /* Node type, These are used by token, node, tree. */
242 #ifdef RE_ENABLE_I18N
245 #endif /* RE_ENABLE_I18N */
247 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
248 when the debugger shows values of this enum type. */
249 #define EPSILON_BIT 8
250 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
251 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
252 OP_ALT = EPSILON_BIT | 2,
253 OP_DUP_ASTERISK = EPSILON_BIT | 3,
254 ANCHOR = EPSILON_BIT | 4,
256 /* Tree type, these are used only by tree. */
260 /* Token type, these are used only by token. */
272 OP_CLOSE_EQUIV_CLASS,
283 #ifdef RE_ENABLE_I18N
286 /* Multibyte characters. */
289 /* Collating symbols. */
294 /* Equivalence classes. */
296 int32_t *equiv_classes;
299 /* Range expressions. */
301 uint32_t *range_starts;
302 uint32_t *range_ends;
303 # else /* not _LIBC */
304 wchar_t *range_starts;
306 # endif /* not _LIBC */
308 /* Character classes. */
309 wctype_t *char_classes;
311 /* If this character set is the non-matching list. */
312 unsigned int non_match : 1;
314 /* # of multibyte characters. */
317 /* # of collating symbols. */
320 /* # of equivalence classes. */
323 /* # of range expressions. */
326 /* # of character classes. */
329 #endif /* RE_ENABLE_I18N */
335 unsigned char c; /* for CHARACTER */
336 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
337 #ifdef RE_ENABLE_I18N
338 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
339 #endif /* RE_ENABLE_I18N */
340 Idx idx; /* for BACK_REF */
341 re_context_type ctx_type; /* for ANCHOR */
343 #if __GNUC__ >= 2 && !defined __STRICT_ANSI__
344 re_token_type_t type : 8;
346 re_token_type_t type;
348 unsigned int constraint : 10; /* context constraint */
349 unsigned int duplicated : 1;
350 unsigned int opt_subexp : 1;
351 #ifdef RE_ENABLE_I18N
352 unsigned int accept_mb : 1;
353 /* These 2 bits can be moved into the union if needed (e.g. if running out
354 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
355 unsigned int mb_partial : 1;
357 unsigned int word_char : 1;
360 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
364 /* Indicate the raw buffer which is the original string passed as an
365 argument of regexec(), re_search(), etc.. */
366 const unsigned char *raw_mbs;
367 /* Store the multibyte string. In case of "case insensitive mode" like
368 REG_ICASE, upper cases of the string are stored, otherwise MBS points
369 the same address that RAW_MBS points. */
371 #ifdef RE_ENABLE_I18N
372 /* Store the wide character string which is corresponding to MBS. */
377 /* Index in RAW_MBS. Each character mbs[i] corresponds to
378 raw_mbs[raw_mbs_idx + i]. */
380 /* The length of the valid characters in the buffers. */
382 /* The corresponding number of bytes in raw_mbs array. */
384 /* The length of the buffers MBS and WCS. */
386 /* The index in MBS, which is updated by re_string_fetch_byte. */
388 /* length of RAW_MBS array. */
390 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
392 /* End of the buffer may be shorter than its length in the cases such
393 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
396 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
399 /* The context of mbs[0]. We store the context independently, since
400 the context of mbs[0] may be different from raw_mbs[0], which is
401 the beginning of the input string. */
402 unsigned int tip_context;
403 /* The translation passed as a part of an argument of re_compile_pattern. */
404 RE_TRANSLATE_TYPE trans;
405 /* Copy of re_dfa_t's word_char. */
406 re_const_bitset_ptr_t word_char;
407 /* true if REG_ICASE. */
409 unsigned char is_utf8;
410 unsigned char map_notascii;
411 unsigned char mbs_allocated;
412 unsigned char offsets_needed;
413 unsigned char newline_anchor;
414 unsigned char word_ops_used;
417 typedef struct re_string_t re_string_t;
421 typedef struct re_dfa_t re_dfa_t;
424 # define internal_function
428 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
431 # ifdef RE_ENABLE_I18N
432 static void build_wcs_buffer (re_string_t *pstr) internal_function;
433 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
435 # endif /* RE_ENABLE_I18N */
436 static void build_upper_buffer (re_string_t *pstr) internal_function;
437 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
438 static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
440 internal_function __attribute__ ((pure));
442 #define re_string_peek_byte(pstr, offset) \
443 ((pstr)->mbs[(pstr)->cur_idx + offset])
444 #define re_string_fetch_byte(pstr) \
445 ((pstr)->mbs[(pstr)->cur_idx++])
446 #define re_string_first_byte(pstr, idx) \
447 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
448 #define re_string_is_single_byte_char(pstr, idx) \
449 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
450 || (pstr)->wcs[(idx) + 1] != WEOF))
451 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
452 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
453 #define re_string_get_buffer(pstr) ((pstr)->mbs)
454 #define re_string_length(pstr) ((pstr)->len)
455 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
456 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
457 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
459 #if defined _LIBC || HAVE_ALLOCA
465 /* The OS usually guarantees only one guard page at the bottom of the stack,
466 and a page size can be as small as 4096 bytes. So we cannot safely
467 allocate anything larger than 4096 bytes. Also care for the possibility
468 of a few compiler-allocated temporary stack slots. */
469 # define __libc_use_alloca(n) ((n) < 4032)
471 /* alloca is implemented with malloc, so just use malloc. */
472 # define __libc_use_alloca(n) 0
474 # define alloca(n) malloc (n)
479 # define MALLOC_0_IS_NONNULL 1
480 #elif !defined MALLOC_0_IS_NONNULL
481 # define MALLOC_0_IS_NONNULL 0
485 # define MAX(a,b) ((a) < (b) ? (b) : (a))
488 # define MIN(a,b) ((a) < (b) ? (a) : (b))
491 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
492 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
493 #define re_free(p) free (p)
497 struct bin_tree_t *parent;
498 struct bin_tree_t *left;
499 struct bin_tree_t *right;
500 struct bin_tree_t *first;
501 struct bin_tree_t *next;
505 /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
506 Otherwise 'type' indicate the type of this node. */
509 typedef struct bin_tree_t bin_tree_t;
511 #define BIN_TREE_STORAGE_SIZE \
512 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
514 struct bin_tree_storage_t
516 struct bin_tree_storage_t *next;
517 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
519 typedef struct bin_tree_storage_t bin_tree_storage_t;
521 #define CONTEXT_WORD 1
522 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
523 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
524 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
526 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
527 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
528 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
529 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
530 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
532 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
533 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
534 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
535 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
537 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
538 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
539 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
540 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
541 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
543 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
544 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
545 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
546 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
547 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
553 re_node_set non_eps_nodes;
554 re_node_set inveclosure;
555 re_node_set *entrance_nodes;
556 struct re_dfastate_t **trtable, **word_trtable;
557 unsigned int context : 4;
558 unsigned int halt : 1;
559 /* If this state can accept "multi byte".
560 Note that we refer to multibyte characters, and multi character
561 collating elements as "multi byte". */
562 unsigned int accept_mb : 1;
563 /* If this state has backreference node(s). */
564 unsigned int has_backref : 1;
565 unsigned int has_constraint : 1;
567 typedef struct re_dfastate_t re_dfastate_t;
569 struct re_state_table_entry
573 re_dfastate_t **array;
576 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
582 re_dfastate_t **array;
585 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
590 Idx str_idx; /* The position NODE match at. */
592 } re_sub_match_last_t;
594 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
595 And information about the node, whose type is OP_CLOSE_SUBEXP,
596 corresponding to NODE is stored in LASTS. */
603 Idx alasts; /* Allocation size of LASTS. */
604 Idx nlasts; /* The number of LASTS. */
605 re_sub_match_last_t **lasts;
606 } re_sub_match_top_t;
608 struct re_backref_cache_entry
616 unsigned short int eps_reachable_subexps_map;
621 /* The string object corresponding to the input string. */
623 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
624 const re_dfa_t *const dfa;
628 /* EFLAGS of the argument of regexec. */
630 /* Where the matching ends. */
633 /* The state log used by the matcher. */
634 re_dfastate_t **state_log;
636 /* Back reference cache. */
639 struct re_backref_cache_entry *bkref_ents;
643 re_sub_match_top_t **sub_tops;
644 } re_match_context_t;
648 re_dfastate_t **sifted_states;
649 re_dfastate_t **limited_states;
655 struct re_fail_stack_ent_t
660 re_node_set eps_via_nodes;
663 struct re_fail_stack_t
667 struct re_fail_stack_ent_t *stack;
678 re_node_set *eclosures;
679 re_node_set *inveclosures;
680 struct re_state_table_entry *state_table;
681 re_dfastate_t *init_state;
682 re_dfastate_t *init_state_word;
683 re_dfastate_t *init_state_nl;
684 re_dfastate_t *init_state_begbuf;
685 bin_tree_t *str_tree;
686 bin_tree_storage_t *str_tree_storage;
687 re_bitset_ptr_t sb_char;
688 int str_tree_storage_idx;
690 /* number of subexpressions 're_nsub' is in regex_t. */
691 re_hashval_t state_hash_mask;
693 Idx nbackref; /* The number of backreference in this dfa. */
695 /* Bitmap expressing which backreference is used. */
696 bitset_word_t used_bkref_map;
697 bitset_word_t completed_bkref_map;
699 unsigned int has_plural_match : 1;
700 /* If this dfa has "multibyte node", which is a backreference or
701 a node which can accept multibyte character or multi character
702 collating element. */
703 unsigned int has_mb_node : 1;
704 unsigned int is_utf8 : 1;
705 unsigned int map_notascii : 1;
706 unsigned int word_ops_used : 1;
717 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
718 #define re_node_set_remove(set,id) \
719 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
720 #define re_node_set_empty(p) ((p)->nelem = 0)
721 #define re_node_set_free(set) re_free ((set)->elems)
735 bracket_elem_type type;
745 /* Functions for bitset_t operation. */
748 bitset_set (bitset_t set, Idx i)
750 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
754 bitset_clear (bitset_t set, Idx i)
756 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
760 bitset_contain (const bitset_t set, Idx i)
762 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
766 bitset_empty (bitset_t set)
768 memset (set, '\0', sizeof (bitset_t));
772 bitset_set_all (bitset_t set)
774 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
775 if (SBC_MAX % BITSET_WORD_BITS != 0)
776 set[BITSET_WORDS - 1] =
777 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
781 bitset_copy (bitset_t dest, const bitset_t src)
783 memcpy (dest, src, sizeof (bitset_t));
786 static void __attribute__ ((unused))
787 bitset_not (bitset_t set)
790 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
791 set[bitset_i] = ~set[bitset_i];
792 if (SBC_MAX % BITSET_WORD_BITS != 0)
793 set[BITSET_WORDS - 1] =
794 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
795 & ~set[BITSET_WORDS - 1]);
798 static void __attribute__ ((unused))
799 bitset_merge (bitset_t dest, const bitset_t src)
802 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
803 dest[bitset_i] |= src[bitset_i];
806 static void __attribute__ ((unused))
807 bitset_mask (bitset_t dest, const bitset_t src)
810 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
811 dest[bitset_i] &= src[bitset_i];
814 #ifdef RE_ENABLE_I18N
815 /* Functions for re_string. */
817 internal_function __attribute__ ((pure, unused))
818 re_string_char_size_at (const re_string_t *pstr, Idx idx)
821 if (pstr->mb_cur_max == 1)
823 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
824 if (pstr->wcs[idx + byte_idx] != WEOF)
830 internal_function __attribute__ ((pure, unused))
831 re_string_wchar_at (const re_string_t *pstr, Idx idx)
833 if (pstr->mb_cur_max == 1)
834 return (wint_t) pstr->mbs[idx];
835 return (wint_t) pstr->wcs[idx];
840 internal_function __attribute__ ((pure, unused))
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;
846 # include <locale/weight.h>
847 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
851 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
852 extra = (const unsigned char *)
853 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
854 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
855 _NL_COLLATE_INDIRECTMB);
857 findidx (&p, pstr->len - idx);
858 return p - pstr->mbs - idx;
865 #endif /* RE_ENABLE_I18N */
867 #ifndef __GNUC_PREREQ
868 # if defined __GNUC__ && defined __GNUC_MINOR__
869 # define __GNUC_PREREQ(maj, min) \
870 ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
872 # define __GNUC_PREREQ(maj, min) 0
876 #if __GNUC_PREREQ (3,4)
877 # undef __attribute_warn_unused_result__
878 # define __attribute_warn_unused_result__ \
879 __attribute__ ((__warn_unused_result__))
881 # define __attribute_warn_unused_result__ /* empty */
884 #endif /* _REGEX_INTERNAL_H */