1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002-2021 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 <https://www.gnu.org/licenses/>. */
20 #ifndef _REGEX_INTERNAL_H
21 #define _REGEX_INTERNAL_H 1
38 #if defined DEBUG && DEBUG != 0
40 # define DEBUG_ASSERT(x) assert (x)
42 # define DEBUG_ASSERT(x) assume (x)
46 # include <libc-lock.h>
47 # define lock_define(name) __libc_lock_define (, name)
48 # define lock_init(lock) (__libc_lock_init (lock), 0)
49 # define lock_fini(lock) ((void) 0)
50 # define lock_lock(lock) __libc_lock_lock (lock)
51 # define lock_unlock(lock) __libc_lock_unlock (lock)
52 #elif defined GNULIB_LOCK && !defined USE_UNLOCKED_IO
53 # include "glthread/lock.h"
54 # define lock_define(name) gl_lock_define (, name)
55 # define lock_init(lock) glthread_lock_init (&(lock))
56 # define lock_fini(lock) glthread_lock_destroy (&(lock))
57 # define lock_lock(lock) glthread_lock_lock (&(lock))
58 # define lock_unlock(lock) glthread_lock_unlock (&(lock))
59 #elif defined GNULIB_PTHREAD && !defined USE_UNLOCKED_IO
61 # define lock_define(name) pthread_mutex_t name;
62 # define lock_init(lock) pthread_mutex_init (&(lock), 0)
63 # define lock_fini(lock) pthread_mutex_destroy (&(lock))
64 # define lock_lock(lock) pthread_mutex_lock (&(lock))
65 # define lock_unlock(lock) pthread_mutex_unlock (&(lock))
67 # define lock_define(name)
68 # define lock_init(lock) 0
69 # define lock_fini(lock) ((void) 0)
70 /* The 'dfa' avoids an "unused variable 'dfa'" warning from GCC. */
71 # define lock_lock(lock) ((void) dfa)
72 # define lock_unlock(lock) ((void) 0)
75 /* In case that the system doesn't have isblank(). */
76 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
77 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
81 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
82 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
83 # include <locale/localeinfo.h>
84 # include <locale/coll-lookup.h>
88 /* This is for other GNU distributions with internationalized messages. */
89 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
93 # define gettext(msgid) \
94 __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
98 # define gettext(msgid) (msgid)
102 /* This define is so xgettext can find the internationalizable
104 # define gettext_noop(String) String
107 #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE) || _LIBC
108 # define RE_ENABLE_I18N
111 /* Number of ASCII characters. */
112 #define ASCII_CHARS 0x80
114 /* Number of single byte characters. */
115 #define SBC_MAX (UCHAR_MAX + 1)
117 #define COLL_ELEM_LEN_MAX 8
119 /* The character which represents newline. */
120 #define NEWLINE_CHAR '\n'
121 #define WIDE_NEWLINE_CHAR L'\n'
123 /* Rename to standard API for using out of glibc. */
130 # define __wctype wctype
131 # define __iswalnum iswalnum
132 # define __iswctype iswctype
133 # define __towlower towlower
134 # define __towupper towupper
135 # define __btowc btowc
136 # define __mbrtowc mbrtowc
137 # define __wcrtomb wcrtomb
138 # define __regfree regfree
139 #endif /* not _LIBC */
142 # define SSIZE_MAX ((ssize_t) (SIZE_MAX / 2))
145 # define ULONG_WIDTH REGEX_UINTEGER_WIDTH (ULONG_MAX)
146 /* The number of usable bits in an unsigned integer type with maximum
147 value MAX, as an int expression suitable in #if. Cover all known
148 practical hosts. This implementation exploits the fact that MAX is
149 1 less than a power of 2, and merely counts the number of 1 bits in
150 MAX; "COBn" means "count the number of 1 bits in the low-order n bits". */
151 # define REGEX_UINTEGER_WIDTH(max) REGEX_COB128 (max)
152 # define REGEX_COB128(n) (REGEX_COB64 ((n) >> 31 >> 31 >> 2) + REGEX_COB64 (n))
153 # define REGEX_COB64(n) (REGEX_COB32 ((n) >> 31 >> 1) + REGEX_COB32 (n))
154 # define REGEX_COB32(n) (REGEX_COB16 ((n) >> 16) + REGEX_COB16 (n))
155 # define REGEX_COB16(n) (REGEX_COB8 ((n) >> 8) + REGEX_COB8 (n))
156 # define REGEX_COB8(n) (REGEX_COB4 ((n) >> 4) + REGEX_COB4 (n))
157 # define REGEX_COB4(n) (!!((n) & 8) + !!((n) & 4) + !!((n) & 2) + ((n) & 1))
158 # if ULONG_MAX / 2 + 1 != 1ul << (ULONG_WIDTH - 1)
159 # error "ULONG_MAX out of range"
163 /* The type of indexes into strings. This is signed, not size_t,
164 since the API requires indexes to fit in regoff_t anyway, and using
165 signed integers makes the code a bit smaller and presumably faster.
166 The traditional GNU regex implementation uses int for indexes.
167 The POSIX-compatible implementation uses a possibly-wider type.
168 The name 'Idx' is three letters to minimize the hassle of
169 reindenting a lot of regex code that formerly used 'int'. */
170 typedef regoff_t Idx
;
171 #ifdef _REGEX_LARGE_OFFSETS
172 # define IDX_MAX SSIZE_MAX
174 # define IDX_MAX INT_MAX
177 /* A hash value, suitable for computing hash tables. */
178 typedef __re_size_t re_hashval_t
;
180 /* An integer used to represent a set of bits. It must be unsigned,
181 and must be at least as wide as unsigned int. */
182 typedef unsigned long int bitset_word_t
;
183 /* All bits set in a bitset_word_t. */
184 #define BITSET_WORD_MAX ULONG_MAX
185 /* Number of bits in a bitset_word_t. */
186 #define BITSET_WORD_BITS ULONG_WIDTH
188 /* Number of bitset_word_t values in a bitset_t. */
189 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
191 typedef bitset_word_t bitset_t
[BITSET_WORDS
];
192 typedef bitset_word_t
*re_bitset_ptr_t
;
193 typedef const bitset_word_t
*re_const_bitset_ptr_t
;
195 #define PREV_WORD_CONSTRAINT 0x0001
196 #define PREV_NOTWORD_CONSTRAINT 0x0002
197 #define NEXT_WORD_CONSTRAINT 0x0004
198 #define NEXT_NOTWORD_CONSTRAINT 0x0008
199 #define PREV_NEWLINE_CONSTRAINT 0x0010
200 #define NEXT_NEWLINE_CONSTRAINT 0x0020
201 #define PREV_BEGBUF_CONSTRAINT 0x0040
202 #define NEXT_ENDBUF_CONSTRAINT 0x0080
203 #define WORD_DELIM_CONSTRAINT 0x0100
204 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
208 INSIDE_WORD
= PREV_WORD_CONSTRAINT
| NEXT_WORD_CONSTRAINT
,
209 WORD_FIRST
= PREV_NOTWORD_CONSTRAINT
| NEXT_WORD_CONSTRAINT
,
210 WORD_LAST
= PREV_WORD_CONSTRAINT
| NEXT_NOTWORD_CONSTRAINT
,
211 INSIDE_NOTWORD
= PREV_NOTWORD_CONSTRAINT
| NEXT_NOTWORD_CONSTRAINT
,
212 LINE_FIRST
= PREV_NEWLINE_CONSTRAINT
,
213 LINE_LAST
= NEXT_NEWLINE_CONSTRAINT
,
214 BUF_FIRST
= PREV_BEGBUF_CONSTRAINT
,
215 BUF_LAST
= NEXT_ENDBUF_CONSTRAINT
,
216 WORD_DELIM
= WORD_DELIM_CONSTRAINT
,
217 NOT_WORD_DELIM
= NOT_WORD_DELIM_CONSTRAINT
231 /* Node type, These are used by token, node, tree. */
237 #ifdef RE_ENABLE_I18N
240 #endif /* RE_ENABLE_I18N */
242 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
243 when the debugger shows values of this enum type. */
244 #define EPSILON_BIT 8
245 OP_OPEN_SUBEXP
= EPSILON_BIT
| 0,
246 OP_CLOSE_SUBEXP
= EPSILON_BIT
| 1,
247 OP_ALT
= EPSILON_BIT
| 2,
248 OP_DUP_ASTERISK
= EPSILON_BIT
| 3,
249 ANCHOR
= EPSILON_BIT
| 4,
251 /* Tree type, these are used only by tree. */
255 /* Token type, these are used only by token. */
267 OP_CLOSE_EQUIV_CLASS
,
278 #ifdef RE_ENABLE_I18N
281 /* Multibyte characters. */
284 /* Collating symbols. */
289 /* Equivalence classes. */
291 int32_t *equiv_classes
;
294 /* Range expressions. */
296 uint32_t *range_starts
;
297 uint32_t *range_ends
;
298 # else /* not _LIBC */
299 wchar_t *range_starts
;
301 # endif /* not _LIBC */
303 /* Character classes. */
304 wctype_t *char_classes
;
306 /* If this character set is the non-matching list. */
307 unsigned int non_match
: 1;
309 /* # of multibyte characters. */
312 /* # of collating symbols. */
315 /* # of equivalence classes. */
318 /* # of range expressions. */
321 /* # of character classes. */
324 #endif /* RE_ENABLE_I18N */
330 unsigned char c
; /* for CHARACTER */
331 re_bitset_ptr_t sbcset
; /* for SIMPLE_BRACKET */
332 #ifdef RE_ENABLE_I18N
333 re_charset_t
*mbcset
; /* for COMPLEX_BRACKET */
334 #endif /* RE_ENABLE_I18N */
335 Idx idx
; /* for BACK_REF */
336 re_context_type ctx_type
; /* for ANCHOR */
338 #if __GNUC__ >= 2 && !defined __STRICT_ANSI__
339 re_token_type_t type
: 8;
341 re_token_type_t type
;
343 unsigned int constraint
: 10; /* context constraint */
344 unsigned int duplicated
: 1;
345 unsigned int opt_subexp
: 1;
346 #ifdef RE_ENABLE_I18N
347 unsigned int accept_mb
: 1;
348 /* These 2 bits can be moved into the union if needed (e.g. if running out
349 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
350 unsigned int mb_partial
: 1;
352 unsigned int word_char
: 1;
355 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
359 /* Indicate the raw buffer which is the original string passed as an
360 argument of regexec(), re_search(), etc.. */
361 const unsigned char *raw_mbs
;
362 /* Store the multibyte string. In case of "case insensitive mode" like
363 REG_ICASE, upper cases of the string are stored, otherwise MBS points
364 the same address that RAW_MBS points. */
366 #ifdef RE_ENABLE_I18N
367 /* Store the wide character string which is corresponding to MBS. */
372 /* Index in RAW_MBS. Each character mbs[i] corresponds to
373 raw_mbs[raw_mbs_idx + i]. */
375 /* The length of the valid characters in the buffers. */
377 /* The corresponding number of bytes in raw_mbs array. */
379 /* The length of the buffers MBS and WCS. */
381 /* The index in MBS, which is updated by re_string_fetch_byte. */
383 /* length of RAW_MBS array. */
385 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
387 /* End of the buffer may be shorter than its length in the cases such
388 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
391 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
394 /* The context of mbs[0]. We store the context independently, since
395 the context of mbs[0] may be different from raw_mbs[0], which is
396 the beginning of the input string. */
397 unsigned int tip_context
;
398 /* The translation passed as a part of an argument of re_compile_pattern. */
399 RE_TRANSLATE_TYPE trans
;
400 /* Copy of re_dfa_t's word_char. */
401 re_const_bitset_ptr_t word_char
;
402 /* true if REG_ICASE. */
404 unsigned char is_utf8
;
405 unsigned char map_notascii
;
406 unsigned char mbs_allocated
;
407 unsigned char offsets_needed
;
408 unsigned char newline_anchor
;
409 unsigned char word_ops_used
;
412 typedef struct re_string_t re_string_t
;
416 typedef struct re_dfa_t re_dfa_t
;
419 # define IS_IN(libc) false
422 #define re_string_peek_byte(pstr, offset) \
423 ((pstr)->mbs[(pstr)->cur_idx + offset])
424 #define re_string_fetch_byte(pstr) \
425 ((pstr)->mbs[(pstr)->cur_idx++])
426 #define re_string_first_byte(pstr, idx) \
427 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
428 #define re_string_is_single_byte_char(pstr, idx) \
429 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
430 || (pstr)->wcs[(idx) + 1] != WEOF))
431 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
432 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
433 #define re_string_get_buffer(pstr) ((pstr)->mbs)
434 #define re_string_length(pstr) ((pstr)->len)
435 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
436 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
437 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
439 #if defined _LIBC || HAVE_ALLOCA
445 /* The OS usually guarantees only one guard page at the bottom of the stack,
446 and a page size can be as small as 4096 bytes. So we cannot safely
447 allocate anything larger than 4096 bytes. Also care for the possibility
448 of a few compiler-allocated temporary stack slots. */
449 # define __libc_use_alloca(n) ((n) < 4032)
451 /* alloca is implemented with malloc, so just use malloc. */
452 # define __libc_use_alloca(n) 0
454 # define alloca(n) malloc (n)
459 # define MALLOC_0_IS_NONNULL 1
460 #elif !defined MALLOC_0_IS_NONNULL
461 # define MALLOC_0_IS_NONNULL 0
465 # define MAX(a,b) ((a) < (b) ? (b) : (a))
468 # define MIN(a,b) ((a) < (b) ? (a) : (b))
471 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
472 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
473 #define re_free(p) free (p)
477 struct bin_tree_t
*parent
;
478 struct bin_tree_t
*left
;
479 struct bin_tree_t
*right
;
480 struct bin_tree_t
*first
;
481 struct bin_tree_t
*next
;
485 /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
486 Otherwise 'type' indicate the type of this node. */
489 typedef struct bin_tree_t bin_tree_t
;
491 #define BIN_TREE_STORAGE_SIZE \
492 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
494 struct bin_tree_storage_t
496 struct bin_tree_storage_t
*next
;
497 bin_tree_t data
[BIN_TREE_STORAGE_SIZE
];
499 typedef struct bin_tree_storage_t bin_tree_storage_t
;
501 #define CONTEXT_WORD 1
502 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
503 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
504 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
506 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
507 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
508 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
509 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
510 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
512 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
513 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
514 #define IS_WIDE_WORD_CHAR(ch) (__iswalnum (ch) || (ch) == L'_')
515 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
517 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
518 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
519 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
520 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
521 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
523 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
524 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
525 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
526 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
527 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
533 re_node_set non_eps_nodes
;
534 re_node_set inveclosure
;
535 re_node_set
*entrance_nodes
;
536 struct re_dfastate_t
**trtable
, **word_trtable
;
537 unsigned int context
: 4;
538 unsigned int halt
: 1;
539 /* If this state can accept "multi byte".
540 Note that we refer to multibyte characters, and multi character
541 collating elements as "multi byte". */
542 unsigned int accept_mb
: 1;
543 /* If this state has backreference node(s). */
544 unsigned int has_backref
: 1;
545 unsigned int has_constraint
: 1;
547 typedef struct re_dfastate_t re_dfastate_t
;
549 struct re_state_table_entry
553 re_dfastate_t
**array
;
556 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
562 re_dfastate_t
**array
;
565 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
570 Idx str_idx
; /* The position NODE match at. */
572 } re_sub_match_last_t
;
574 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
575 And information about the node, whose type is OP_CLOSE_SUBEXP,
576 corresponding to NODE is stored in LASTS. */
583 Idx alasts
; /* Allocation size of LASTS. */
584 Idx nlasts
; /* The number of LASTS. */
585 re_sub_match_last_t
**lasts
;
586 } re_sub_match_top_t
;
588 struct re_backref_cache_entry
594 bitset_word_t eps_reachable_subexps_map
;
600 /* The string object corresponding to the input string. */
602 const re_dfa_t
*const dfa
;
603 /* EFLAGS of the argument of regexec. */
605 /* Where the matching ends. */
608 /* The state log used by the matcher. */
609 re_dfastate_t
**state_log
;
611 /* Back reference cache. */
614 struct re_backref_cache_entry
*bkref_ents
;
618 re_sub_match_top_t
**sub_tops
;
619 } re_match_context_t
;
623 re_dfastate_t
**sifted_states
;
624 re_dfastate_t
**limited_states
;
630 struct re_fail_stack_ent_t
635 re_node_set eps_via_nodes
;
638 struct re_fail_stack_t
642 struct re_fail_stack_ent_t
*stack
;
653 re_node_set
*eclosures
;
654 re_node_set
*inveclosures
;
655 struct re_state_table_entry
*state_table
;
656 re_dfastate_t
*init_state
;
657 re_dfastate_t
*init_state_word
;
658 re_dfastate_t
*init_state_nl
;
659 re_dfastate_t
*init_state_begbuf
;
660 bin_tree_t
*str_tree
;
661 bin_tree_storage_t
*str_tree_storage
;
662 re_bitset_ptr_t sb_char
;
663 int str_tree_storage_idx
;
665 /* number of subexpressions 're_nsub' is in regex_t. */
666 re_hashval_t state_hash_mask
;
668 Idx nbackref
; /* The number of backreference in this dfa. */
670 /* Bitmap expressing which backreference is used. */
671 bitset_word_t used_bkref_map
;
672 bitset_word_t completed_bkref_map
;
674 unsigned int has_plural_match
: 1;
675 /* If this dfa has "multibyte node", which is a backreference or
676 a node which can accept multibyte character or multi character
677 collating element. */
678 unsigned int has_mb_node
: 1;
679 unsigned int is_utf8
: 1;
680 unsigned int map_notascii
: 1;
681 unsigned int word_ops_used
: 1;
692 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
693 #define re_node_set_remove(set,id) \
694 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
695 #define re_node_set_empty(p) ((p)->nelem = 0)
696 #define re_node_set_free(set) re_free ((set)->elems)
710 bracket_elem_type type
;
720 /* Functions for bitset_t operation. */
723 bitset_set (bitset_t set
, Idx i
)
725 set
[i
/ BITSET_WORD_BITS
] |= (bitset_word_t
) 1 << i
% BITSET_WORD_BITS
;
729 bitset_clear (bitset_t set
, Idx i
)
731 set
[i
/ BITSET_WORD_BITS
] &= ~ ((bitset_word_t
) 1 << i
% BITSET_WORD_BITS
);
735 bitset_contain (const bitset_t set
, Idx i
)
737 return (set
[i
/ BITSET_WORD_BITS
] >> i
% BITSET_WORD_BITS
) & 1;
741 bitset_empty (bitset_t set
)
743 memset (set
, '\0', sizeof (bitset_t
));
747 bitset_set_all (bitset_t set
)
749 memset (set
, -1, sizeof (bitset_word_t
) * (SBC_MAX
/ BITSET_WORD_BITS
));
750 if (SBC_MAX
% BITSET_WORD_BITS
!= 0)
751 set
[BITSET_WORDS
- 1] =
752 ((bitset_word_t
) 1 << SBC_MAX
% BITSET_WORD_BITS
) - 1;
756 bitset_copy (bitset_t dest
, const bitset_t src
)
758 memcpy (dest
, src
, sizeof (bitset_t
));
762 bitset_not (bitset_t set
)
765 for (bitset_i
= 0; bitset_i
< SBC_MAX
/ BITSET_WORD_BITS
; ++bitset_i
)
766 set
[bitset_i
] = ~set
[bitset_i
];
767 if (SBC_MAX
% BITSET_WORD_BITS
!= 0)
768 set
[BITSET_WORDS
- 1] =
769 ((((bitset_word_t
) 1 << SBC_MAX
% BITSET_WORD_BITS
) - 1)
770 & ~set
[BITSET_WORDS
- 1]);
774 bitset_merge (bitset_t dest
, const bitset_t src
)
777 for (bitset_i
= 0; bitset_i
< BITSET_WORDS
; ++bitset_i
)
778 dest
[bitset_i
] |= src
[bitset_i
];
782 bitset_mask (bitset_t dest
, const bitset_t src
)
785 for (bitset_i
= 0; bitset_i
< BITSET_WORDS
; ++bitset_i
)
786 dest
[bitset_i
] &= src
[bitset_i
];
789 #ifdef RE_ENABLE_I18N
790 /* Functions for re_string. */
792 __attribute__ ((pure
, unused
))
793 re_string_char_size_at (const re_string_t
*pstr
, Idx idx
)
796 if (pstr
->mb_cur_max
== 1)
798 for (byte_idx
= 1; idx
+ byte_idx
< pstr
->valid_len
; ++byte_idx
)
799 if (pstr
->wcs
[idx
+ byte_idx
] != WEOF
)
805 __attribute__ ((pure
, unused
))
806 re_string_wchar_at (const re_string_t
*pstr
, Idx idx
)
808 if (pstr
->mb_cur_max
== 1)
809 return (wint_t) pstr
->mbs
[idx
];
810 return (wint_t) pstr
->wcs
[idx
];
814 # include <locale/weight.h>
818 __attribute__ ((pure
, unused
))
819 re_string_elem_size_at (const re_string_t
*pstr
, Idx idx
)
822 const unsigned char *p
, *extra
;
823 const int32_t *table
, *indirect
;
824 uint_fast32_t nrules
= _NL_CURRENT_WORD (LC_COLLATE
, _NL_COLLATE_NRULES
);
828 table
= (const int32_t *) _NL_CURRENT (LC_COLLATE
, _NL_COLLATE_TABLEMB
);
829 extra
= (const unsigned char *)
830 _NL_CURRENT (LC_COLLATE
, _NL_COLLATE_EXTRAMB
);
831 indirect
= (const int32_t *) _NL_CURRENT (LC_COLLATE
,
832 _NL_COLLATE_INDIRECTMB
);
834 findidx (table
, indirect
, extra
, &p
, pstr
->len
- idx
);
835 return p
- pstr
->mbs
- idx
;
841 #endif /* RE_ENABLE_I18N */
845 # define FALLTHROUGH ((void) 0)
847 # define FALLTHROUGH __attribute__ ((__fallthrough__))
851 #endif /* _REGEX_INTERNAL_H */