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 # define lock_define(name) gl_lock_define (, name)
46 # define lock_init(lock) glthread_lock_init (&(lock))
47 # define lock_fini(lock) glthread_lock_destroy (&(lock))
48 # define lock_lock(lock) glthread_lock_lock (&(lock))
49 # define lock_unlock(lock) glthread_lock_unlock (&(lock))
50 #elif defined GNULIB_PTHREAD
52 # define lock_define(name) pthread_mutex_t name;
53 # define lock_init(lock) pthread_mutex_init (&(lock), 0)
54 # define lock_fini(lock) pthread_mutex_destroy (&(lock))
55 # define lock_lock(lock) pthread_mutex_lock (&(lock))
56 # define lock_unlock(lock) pthread_mutex_unlock (&(lock))
58 # define lock_define(name)
59 # define lock_init(lock) 0
60 # define lock_fini(lock) 0
61 # define lock_lock(lock) ((void) 0)
62 # define lock_unlock(lock) ((void) 0)
65 /* In case that the system doesn't have isblank(). */
66 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
67 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
71 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
72 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
73 # include <locale/localeinfo.h>
74 # include <locale/elem-hash.h>
75 # include <locale/coll-lookup.h>
79 /* This is for other GNU distributions with internationalized messages. */
80 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
84 # define gettext(msgid) \
85 __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
88 # define gettext(msgid) (msgid)
92 /* This define is so xgettext can find the internationalizable
94 # define gettext_noop(String) String
97 #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE && HAVE_WCSCOLL) || _LIBC
98 # define RE_ENABLE_I18N
102 # define BE(expr, val) __builtin_expect (expr, val)
104 # define BE(expr, val) (expr)
107 /* Number of ASCII characters. */
108 #define ASCII_CHARS 0x80
110 /* Number of single byte characters. */
111 #define SBC_MAX (UCHAR_MAX + 1)
113 #define COLL_ELEM_LEN_MAX 8
115 /* The character which represents newline. */
116 #define NEWLINE_CHAR '\n'
117 #define WIDE_NEWLINE_CHAR L'\n'
119 /* Rename to standard API for using out of glibc. */
123 # define __wctype wctype
124 # define __iswctype iswctype
125 # define __btowc btowc
126 # define __mbrtowc mbrtowc
127 # define __wcrtomb wcrtomb
128 # define __regfree regfree
129 # define attribute_hidden
130 #endif /* not _LIBC */
132 #if __GNUC__ < 3 + (__GNUC_MINOR__ < 1)
133 # define __attribute__(arg)
136 typedef __re_idx_t Idx;
137 #ifdef _REGEX_LARGE_OFFSETS
138 # define IDX_MAX (SIZE_MAX - 2)
140 # define IDX_MAX INT_MAX
143 /* Special return value for failure to match. */
144 #define REG_MISSING ((Idx) -1)
146 /* Special return value for internal error. */
147 #define REG_ERROR ((Idx) -2)
149 /* Test whether N is a valid index, and is not one of the above. */
150 #ifdef _REGEX_LARGE_OFFSETS
151 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
153 # define REG_VALID_INDEX(n) (0 <= (n))
156 /* Test whether N is a valid nonzero index. */
157 #ifdef _REGEX_LARGE_OFFSETS
158 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
160 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
163 /* A hash value, suitable for computing hash tables. */
164 typedef __re_size_t re_hashval_t;
166 /* An integer used to represent a set of bits. It must be unsigned,
167 and must be at least as wide as unsigned int. */
168 typedef unsigned long int bitset_word_t;
169 /* All bits set in a bitset_word_t. */
170 #define BITSET_WORD_MAX ULONG_MAX
172 /* Number of bits in a bitset_word_t. For portability to hosts with
173 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
174 instead, deduce it directly from BITSET_WORD_MAX. Avoid
175 greater-than-32-bit integers and unconditional shifts by more than
176 31 bits, as they're not portable. */
177 #if BITSET_WORD_MAX == 0xffffffffUL
178 # define BITSET_WORD_BITS 32
179 #elif BITSET_WORD_MAX >> 31 >> 4 == 1
180 # define BITSET_WORD_BITS 36
181 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
182 # define BITSET_WORD_BITS 48
183 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
184 # define BITSET_WORD_BITS 60
185 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
186 # define BITSET_WORD_BITS 64
187 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
188 # define BITSET_WORD_BITS 72
189 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
190 # define BITSET_WORD_BITS 128
191 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
192 # define BITSET_WORD_BITS 256
193 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
194 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
195 # if BITSET_WORD_BITS <= SBC_MAX
196 # error "Invalid SBC_MAX"
199 # error "Add case for new bitset_word_t size"
202 /* Number of bitset_word_t values in a bitset_t. */
203 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
205 typedef bitset_word_t bitset_t[BITSET_WORDS];
206 typedef bitset_word_t *re_bitset_ptr_t;
207 typedef const bitset_word_t *re_const_bitset_ptr_t;
209 #define PREV_WORD_CONSTRAINT 0x0001
210 #define PREV_NOTWORD_CONSTRAINT 0x0002
211 #define NEXT_WORD_CONSTRAINT 0x0004
212 #define NEXT_NOTWORD_CONSTRAINT 0x0008
213 #define PREV_NEWLINE_CONSTRAINT 0x0010
214 #define NEXT_NEWLINE_CONSTRAINT 0x0020
215 #define PREV_BEGBUF_CONSTRAINT 0x0040
216 #define NEXT_ENDBUF_CONSTRAINT 0x0080
217 #define WORD_DELIM_CONSTRAINT 0x0100
218 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
222 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
223 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
224 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
225 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
226 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
227 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
228 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
229 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
230 WORD_DELIM = WORD_DELIM_CONSTRAINT,
231 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
245 /* Node type, These are used by token, node, tree. */
251 #ifdef RE_ENABLE_I18N
254 #endif /* RE_ENABLE_I18N */
256 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
257 when the debugger shows values of this enum type. */
258 #define EPSILON_BIT 8
259 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
260 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
261 OP_ALT = EPSILON_BIT | 2,
262 OP_DUP_ASTERISK = EPSILON_BIT | 3,
263 ANCHOR = EPSILON_BIT | 4,
265 /* Tree type, these are used only by tree. */
269 /* Token type, these are used only by token. */
281 OP_CLOSE_EQUIV_CLASS,
292 #ifdef RE_ENABLE_I18N
295 /* Multibyte characters. */
298 /* Collating symbols. */
303 /* Equivalence classes. */
305 int32_t *equiv_classes;
308 /* Range expressions. */
310 uint32_t *range_starts;
311 uint32_t *range_ends;
312 # else /* not _LIBC */
313 wchar_t *range_starts;
315 # endif /* not _LIBC */
317 /* Character classes. */
318 wctype_t *char_classes;
320 /* If this character set is the non-matching list. */
321 unsigned int non_match : 1;
323 /* # of multibyte characters. */
326 /* # of collating symbols. */
329 /* # of equivalence classes. */
332 /* # of range expressions. */
335 /* # of character classes. */
338 #endif /* RE_ENABLE_I18N */
344 unsigned char c; /* for CHARACTER */
345 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
346 #ifdef RE_ENABLE_I18N
347 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
348 #endif /* RE_ENABLE_I18N */
349 Idx idx; /* for BACK_REF */
350 re_context_type ctx_type; /* for ANCHOR */
352 #if __GNUC__ >= 2 && !defined __STRICT_ANSI__
353 re_token_type_t type : 8;
355 re_token_type_t type;
357 unsigned int constraint : 10; /* context constraint */
358 unsigned int duplicated : 1;
359 unsigned int opt_subexp : 1;
360 #ifdef RE_ENABLE_I18N
361 unsigned int accept_mb : 1;
362 /* These 2 bits can be moved into the union if needed (e.g. if running out
363 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
364 unsigned int mb_partial : 1;
366 unsigned int word_char : 1;
369 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
373 /* Indicate the raw buffer which is the original string passed as an
374 argument of regexec(), re_search(), etc.. */
375 const unsigned char *raw_mbs;
376 /* Store the multibyte string. In case of "case insensitive mode" like
377 REG_ICASE, upper cases of the string are stored, otherwise MBS points
378 the same address that RAW_MBS points. */
380 #ifdef RE_ENABLE_I18N
381 /* Store the wide character string which is corresponding to MBS. */
386 /* Index in RAW_MBS. Each character mbs[i] corresponds to
387 raw_mbs[raw_mbs_idx + i]. */
389 /* The length of the valid characters in the buffers. */
391 /* The corresponding number of bytes in raw_mbs array. */
393 /* The length of the buffers MBS and WCS. */
395 /* The index in MBS, which is updated by re_string_fetch_byte. */
397 /* length of RAW_MBS array. */
399 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
401 /* End of the buffer may be shorter than its length in the cases such
402 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
405 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
408 /* The context of mbs[0]. We store the context independently, since
409 the context of mbs[0] may be different from raw_mbs[0], which is
410 the beginning of the input string. */
411 unsigned int tip_context;
412 /* The translation passed as a part of an argument of re_compile_pattern. */
413 RE_TRANSLATE_TYPE trans;
414 /* Copy of re_dfa_t's word_char. */
415 re_const_bitset_ptr_t word_char;
416 /* true if REG_ICASE. */
418 unsigned char is_utf8;
419 unsigned char map_notascii;
420 unsigned char mbs_allocated;
421 unsigned char offsets_needed;
422 unsigned char newline_anchor;
423 unsigned char word_ops_used;
426 typedef struct re_string_t re_string_t;
430 typedef struct re_dfa_t re_dfa_t;
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));
451 #define re_string_peek_byte(pstr, offset) \
452 ((pstr)->mbs[(pstr)->cur_idx + offset])
453 #define re_string_fetch_byte(pstr) \
454 ((pstr)->mbs[(pstr)->cur_idx++])
455 #define re_string_first_byte(pstr, idx) \
456 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
457 #define re_string_is_single_byte_char(pstr, idx) \
458 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
459 || (pstr)->wcs[(idx) + 1] != WEOF))
460 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
461 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
462 #define re_string_get_buffer(pstr) ((pstr)->mbs)
463 #define re_string_length(pstr) ((pstr)->len)
464 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
465 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
466 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
468 #if defined _LIBC || HAVE_ALLOCA
474 /* The OS usually guarantees only one guard page at the bottom of the stack,
475 and a page size can be as small as 4096 bytes. So we cannot safely
476 allocate anything larger than 4096 bytes. Also care for the possibility
477 of a few compiler-allocated temporary stack slots. */
478 # define __libc_use_alloca(n) ((n) < 4032)
480 /* alloca is implemented with malloc, so just use malloc. */
481 # define __libc_use_alloca(n) 0
483 # define alloca(n) malloc (n)
488 # define MALLOC_0_IS_NONNULL 1
489 #elif !defined MALLOC_0_IS_NONNULL
490 # define MALLOC_0_IS_NONNULL 0
494 # define MAX(a,b) ((a) < (b) ? (b) : (a))
497 # define MIN(a,b) ((a) < (b) ? (a) : (b))
500 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
501 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
502 #define re_free(p) free (p)
506 struct bin_tree_t *parent;
507 struct bin_tree_t *left;
508 struct bin_tree_t *right;
509 struct bin_tree_t *first;
510 struct bin_tree_t *next;
514 /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
515 Otherwise 'type' indicate the type of this node. */
518 typedef struct bin_tree_t bin_tree_t;
520 #define BIN_TREE_STORAGE_SIZE \
521 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
523 struct bin_tree_storage_t
525 struct bin_tree_storage_t *next;
526 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
528 typedef struct bin_tree_storage_t bin_tree_storage_t;
530 #define CONTEXT_WORD 1
531 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
532 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
533 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
535 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
536 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
537 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
538 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
539 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
541 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
542 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
543 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
544 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
546 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
547 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
548 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
549 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
550 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
552 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
553 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
554 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
555 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
556 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
562 re_node_set non_eps_nodes;
563 re_node_set inveclosure;
564 re_node_set *entrance_nodes;
565 struct re_dfastate_t **trtable, **word_trtable;
566 unsigned int context : 4;
567 unsigned int halt : 1;
568 /* If this state can accept "multi byte".
569 Note that we refer to multibyte characters, and multi character
570 collating elements as "multi byte". */
571 unsigned int accept_mb : 1;
572 /* If this state has backreference node(s). */
573 unsigned int has_backref : 1;
574 unsigned int has_constraint : 1;
576 typedef struct re_dfastate_t re_dfastate_t;
578 struct re_state_table_entry
582 re_dfastate_t **array;
585 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
591 re_dfastate_t **array;
594 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
599 Idx str_idx; /* The position NODE match at. */
601 } re_sub_match_last_t;
603 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
604 And information about the node, whose type is OP_CLOSE_SUBEXP,
605 corresponding to NODE is stored in LASTS. */
612 Idx alasts; /* Allocation size of LASTS. */
613 Idx nlasts; /* The number of LASTS. */
614 re_sub_match_last_t **lasts;
615 } re_sub_match_top_t;
617 struct re_backref_cache_entry
625 unsigned short int eps_reachable_subexps_map;
630 /* The string object corresponding to the input string. */
632 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
633 const re_dfa_t *const dfa;
637 /* EFLAGS of the argument of regexec. */
639 /* Where the matching ends. */
642 /* The state log used by the matcher. */
643 re_dfastate_t **state_log;
645 /* Back reference cache. */
648 struct re_backref_cache_entry *bkref_ents;
652 re_sub_match_top_t **sub_tops;
653 } re_match_context_t;
657 re_dfastate_t **sifted_states;
658 re_dfastate_t **limited_states;
664 struct re_fail_stack_ent_t
669 re_node_set eps_via_nodes;
672 struct re_fail_stack_t
676 struct re_fail_stack_ent_t *stack;
687 re_node_set *eclosures;
688 re_node_set *inveclosures;
689 struct re_state_table_entry *state_table;
690 re_dfastate_t *init_state;
691 re_dfastate_t *init_state_word;
692 re_dfastate_t *init_state_nl;
693 re_dfastate_t *init_state_begbuf;
694 bin_tree_t *str_tree;
695 bin_tree_storage_t *str_tree_storage;
696 re_bitset_ptr_t sb_char;
697 int str_tree_storage_idx;
699 /* number of subexpressions 're_nsub' is in regex_t. */
700 re_hashval_t state_hash_mask;
702 Idx nbackref; /* The number of backreference in this dfa. */
704 /* Bitmap expressing which backreference is used. */
705 bitset_word_t used_bkref_map;
706 bitset_word_t completed_bkref_map;
708 unsigned int has_plural_match : 1;
709 /* If this dfa has "multibyte node", which is a backreference or
710 a node which can accept multibyte character or multi character
711 collating element. */
712 unsigned int has_mb_node : 1;
713 unsigned int is_utf8 : 1;
714 unsigned int map_notascii : 1;
715 unsigned int word_ops_used : 1;
726 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
727 #define re_node_set_remove(set,id) \
728 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
729 #define re_node_set_empty(p) ((p)->nelem = 0)
730 #define re_node_set_free(set) re_free ((set)->elems)
744 bracket_elem_type type;
754 /* Functions for bitset_t operation. */
757 bitset_set (bitset_t set, Idx i)
759 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
763 bitset_clear (bitset_t set, Idx i)
765 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
769 bitset_contain (const bitset_t set, Idx i)
771 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
775 bitset_empty (bitset_t set)
777 memset (set, '\0', sizeof (bitset_t));
781 bitset_set_all (bitset_t set)
783 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
784 if (SBC_MAX % BITSET_WORD_BITS != 0)
785 set[BITSET_WORDS - 1] =
786 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
790 bitset_copy (bitset_t dest, const bitset_t src)
792 memcpy (dest, src, sizeof (bitset_t));
795 static void __attribute__ ((unused))
796 bitset_not (bitset_t set)
799 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
800 set[bitset_i] = ~set[bitset_i];
801 if (SBC_MAX % BITSET_WORD_BITS != 0)
802 set[BITSET_WORDS - 1] =
803 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
804 & ~set[BITSET_WORDS - 1]);
807 static void __attribute__ ((unused))
808 bitset_merge (bitset_t dest, const bitset_t src)
811 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
812 dest[bitset_i] |= src[bitset_i];
815 static void __attribute__ ((unused))
816 bitset_mask (bitset_t dest, const bitset_t src)
819 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
820 dest[bitset_i] &= src[bitset_i];
823 #ifdef RE_ENABLE_I18N
824 /* Functions for re_string. */
826 internal_function __attribute__ ((pure, unused))
827 re_string_char_size_at (const re_string_t *pstr, Idx idx)
830 if (pstr->mb_cur_max == 1)
832 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
833 if (pstr->wcs[idx + byte_idx] != WEOF)
839 internal_function __attribute__ ((pure, unused))
840 re_string_wchar_at (const re_string_t *pstr, Idx idx)
842 if (pstr->mb_cur_max == 1)
843 return (wint_t) pstr->mbs[idx];
844 return (wint_t) pstr->wcs[idx];
849 internal_function __attribute__ ((pure, unused))
850 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
853 const unsigned char *p, *extra;
854 const int32_t *table, *indirect;
855 # include <locale/weight.h>
856 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
860 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
861 extra = (const unsigned char *)
862 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
863 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
864 _NL_COLLATE_INDIRECTMB);
866 findidx (&p, pstr->len - idx);
867 return p - pstr->mbs - idx;
874 #endif /* RE_ENABLE_I18N */
876 #ifndef __GNUC_PREREQ
877 # if defined __GNUC__ && defined __GNUC_MINOR__
878 # define __GNUC_PREREQ(maj, min) \
879 ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
881 # define __GNUC_PREREQ(maj, min) 0
885 #if __GNUC_PREREQ (3,4)
886 # undef __attribute_warn_unused_result__
887 # define __attribute_warn_unused_result__ \
888 __attribute__ ((__warn_unused_result__))
890 # define __attribute_warn_unused_result__ /* empty */
893 #endif /* _REGEX_INTERNAL_H */