1
/* Extended regular expression matching and search library.
2
Copyright (C) 2002-2023 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
35
# include <dynarray.h>
41
#if defined DEBUG && DEBUG != 0
43
# define DEBUG_ASSERT(x) assert (x)
45
# define DEBUG_ASSERT(x) assume (x)
49
# include <libc-lock.h>
50
# define lock_define(name) __libc_lock_define (, name)
51
# define lock_init(lock) (__libc_lock_init (lock), 0)
52
# define lock_fini(lock) ((void) 0)
53
# define lock_lock(lock) __libc_lock_lock (lock)
54
# define lock_unlock(lock) __libc_lock_unlock (lock)
55
#elif defined GNULIB_LOCK && !defined GNULIB_REGEX_SINGLE_THREAD
56
# include "glthread/lock.h"
57
# define lock_define(name) gl_lock_define (, name)
58
# define lock_init(lock) glthread_lock_init (&(lock))
59
# define lock_fini(lock) glthread_lock_destroy (&(lock))
60
# define lock_lock(lock) glthread_lock_lock (&(lock))
61
# define lock_unlock(lock) glthread_lock_unlock (&(lock))
62
#elif defined GNULIB_PTHREAD && !defined GNULIB_REGEX_SINGLE_THREAD
64
# define lock_define(name) pthread_mutex_t name;
65
# define lock_init(lock) pthread_mutex_init (&(lock), 0)
66
# define lock_fini(lock) pthread_mutex_destroy (&(lock))
67
# define lock_lock(lock) pthread_mutex_lock (&(lock))
68
# define lock_unlock(lock) pthread_mutex_unlock (&(lock))
70
# define lock_define(name)
71
# define lock_init(lock) 0
72
# define lock_fini(lock) ((void) 0)
73
/* The 'dfa' avoids an "unused variable 'dfa'" warning from GCC. */
74
# define lock_lock(lock) ((void) dfa)
75
# define lock_unlock(lock) ((void) 0)
78
/* In case that the system doesn't have isblank(). */
79
#if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
80
# define isblank(ch) ((ch) == ' ' || (ch) == '\t')
83
/* regex code assumes isascii has its usual numeric meaning,
84
even if the portable character set uses EBCDIC encoding,
85
and even if wint_t is wider than int. */
88
# define isascii(c) (((c) & ~0x7f) == 0)
92
# ifndef _RE_DEFINE_LOCALE_FUNCTIONS
93
# define _RE_DEFINE_LOCALE_FUNCTIONS 1
94
# include <locale/localeinfo.h>
95
# include <locale/coll-lookup.h>
99
/* This is for other GNU distributions with internationalized messages. */
100
#if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
101
# include <libintl.h>
104
# define gettext(msgid) \
105
__dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
109
# define gettext(msgid) (msgid)
113
/* This define is so xgettext can find the internationalizable
115
# define gettext_noop(String) String
118
/* Number of ASCII characters. */
119
#define ASCII_CHARS 0x80
121
/* Number of single byte characters. */
122
#define SBC_MAX (UCHAR_MAX + 1)
124
#define COLL_ELEM_LEN_MAX 8
126
/* The character which represents newline. */
127
#define NEWLINE_CHAR '\n'
128
#define WIDE_NEWLINE_CHAR L'\n'
130
/* Rename to standard API for using out of glibc. */
137
# define __wctype wctype
138
# define __iswalnum iswalnum
139
# define __iswctype iswctype
140
# define __towlower towlower
141
# define __towupper towupper
142
# define __btowc btowc
143
# define __mbrtowc mbrtowc
144
# define __wcrtomb wcrtomb
145
# define __regfree regfree
146
#endif /* not _LIBC */
148
/* Types related to integers. Unless protected by #ifdef _LIBC, the
149
regex code should avoid exact-width types like int32_t and uint64_t
150
as some non-GCC platforms lack them, an issue when this code is
154
# define SSIZE_MAX ((ssize_t) (SIZE_MAX / 2))
157
# define ULONG_WIDTH REGEX_UINTEGER_WIDTH (ULONG_MAX)
158
/* The number of usable bits in an unsigned integer type with maximum
159
value MAX, as an int expression suitable in #if. Cover all known
160
practical hosts. This implementation exploits the fact that MAX is
161
1 less than a power of 2, and merely counts the number of 1 bits in
162
MAX; "COBn" means "count the number of 1 bits in the low-order n bits". */
163
# define REGEX_UINTEGER_WIDTH(max) REGEX_COB128 (max)
164
# define REGEX_COB128(n) (REGEX_COB64 ((n) >> 31 >> 31 >> 2) + REGEX_COB64 (n))
165
# define REGEX_COB64(n) (REGEX_COB32 ((n) >> 31 >> 1) + REGEX_COB32 (n))
166
# define REGEX_COB32(n) (REGEX_COB16 ((n) >> 16) + REGEX_COB16 (n))
167
# define REGEX_COB16(n) (REGEX_COB8 ((n) >> 8) + REGEX_COB8 (n))
168
# define REGEX_COB8(n) (REGEX_COB4 ((n) >> 4) + REGEX_COB4 (n))
169
# define REGEX_COB4(n) (!!((n) & 8) + !!((n) & 4) + !!((n) & 2) + ((n) & 1))
170
# if ULONG_MAX / 2 + 1 != 1ul << (ULONG_WIDTH - 1)
171
# error "ULONG_MAX out of range"
175
/* The type of indexes into strings. This is signed, not size_t,
176
since the API requires indexes to fit in regoff_t anyway, and using
177
signed integers makes the code a bit smaller and presumably faster.
178
The traditional GNU regex implementation uses int for indexes.
179
The POSIX-compatible implementation uses a possibly-wider type.
180
The name 'Idx' is three letters to minimize the hassle of
181
reindenting a lot of regex code that formerly used 'int'. */
182
typedef regoff_t Idx;
183
#ifdef _REGEX_LARGE_OFFSETS
184
# define IDX_MAX SSIZE_MAX
186
# define IDX_MAX INT_MAX
189
/* A hash value, suitable for computing hash tables. */
190
typedef __re_size_t re_hashval_t;
192
/* An integer used to represent a set of bits. It must be unsigned,
193
and must be at least as wide as unsigned int. */
194
typedef unsigned long int bitset_word_t;
195
/* All bits set in a bitset_word_t. */
196
#define BITSET_WORD_MAX ULONG_MAX
197
/* Number of bits in a bitset_word_t. */
198
#define BITSET_WORD_BITS ULONG_WIDTH
200
/* Number of bitset_word_t values in a bitset_t. */
201
#define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
203
typedef bitset_word_t bitset_t[BITSET_WORDS];
204
typedef bitset_word_t *re_bitset_ptr_t;
205
typedef const bitset_word_t *re_const_bitset_ptr_t;
207
#define PREV_WORD_CONSTRAINT 0x0001
208
#define PREV_NOTWORD_CONSTRAINT 0x0002
209
#define NEXT_WORD_CONSTRAINT 0x0004
210
#define NEXT_NOTWORD_CONSTRAINT 0x0008
211
#define PREV_NEWLINE_CONSTRAINT 0x0010
212
#define NEXT_NEWLINE_CONSTRAINT 0x0020
213
#define PREV_BEGBUF_CONSTRAINT 0x0040
214
#define NEXT_ENDBUF_CONSTRAINT 0x0080
215
#define WORD_DELIM_CONSTRAINT 0x0100
216
#define NOT_WORD_DELIM_CONSTRAINT 0x0200
220
INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
221
WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
222
WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
223
INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
224
LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
225
LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
226
BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
227
BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
228
WORD_DELIM = WORD_DELIM_CONSTRAINT,
229
NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
243
/* Node type, These are used by token, node, tree. */
252
/* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
253
when the debugger shows values of this enum type. */
254
#define EPSILON_BIT 8
255
OP_OPEN_SUBEXP = EPSILON_BIT | 0,
256
OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
257
OP_ALT = EPSILON_BIT | 2,
258
OP_DUP_ASTERISK = EPSILON_BIT | 3,
259
ANCHOR = EPSILON_BIT | 4,
261
/* Tree type, these are used only by tree. */
265
/* Token type, these are used only by token. */
277
OP_CLOSE_EQUIV_CLASS,
290
/* Multibyte characters. */
294
/* Collating symbols. */
299
/* Equivalence classes. */
300
int32_t *equiv_classes;
303
/* Range expressions. */
305
uint32_t *range_starts;
306
uint32_t *range_ends;
308
wchar_t *range_starts;
312
/* Character classes. */
313
wctype_t *char_classes;
315
/* If this character set is the non-matching list. */
316
unsigned int non_match : 1;
318
/* # of multibyte characters. */
321
/* # of collating symbols. */
324
/* # of equivalence classes. */
327
/* # of range expressions. */
330
/* # of character classes. */
338
unsigned char c; /* for CHARACTER */
339
re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
340
re_charset_t *mbcset; /* for COMPLEX_BRACKET */
341
Idx idx; /* for BACK_REF */
342
re_context_type ctx_type; /* for ANCHOR */
344
#if (__GNUC__ >= 2 || defined __clang__) && !defined __STRICT_ANSI__
345
re_token_type_t type : 8;
347
re_token_type_t type;
349
unsigned int constraint : 10; /* context constraint */
350
unsigned int duplicated : 1;
351
unsigned int opt_subexp : 1;
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;
356
unsigned int word_char : 1;
359
#define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
363
/* Indicate the raw buffer which is the original string passed as an
364
argument of regexec(), re_search(), etc.. */
365
const unsigned char *raw_mbs;
366
/* Store the multibyte string. In case of "case insensitive mode" like
367
REG_ICASE, upper cases of the string are stored, otherwise MBS points
368
the same address that RAW_MBS points. */
370
/* Store the wide character string which is corresponding to MBS. */
374
/* Index in RAW_MBS. Each character mbs[i] corresponds to
375
raw_mbs[raw_mbs_idx + i]. */
377
/* The length of the valid characters in the buffers. */
379
/* The corresponding number of bytes in raw_mbs array. */
381
/* The length of the buffers MBS and WCS. */
383
/* The index in MBS, which is updated by re_string_fetch_byte. */
385
/* length of RAW_MBS array. */
387
/* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
389
/* End of the buffer may be shorter than its length in the cases such
390
as re_match_2, re_search_2. Then, we use STOP for end of the buffer
393
/* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
396
/* The context of mbs[0]. We store the context independently, since
397
the context of mbs[0] may be different from raw_mbs[0], which is
398
the beginning of the input string. */
399
unsigned int tip_context;
400
/* The translation passed as a part of an argument of re_compile_pattern. */
401
RE_TRANSLATE_TYPE trans;
402
/* Copy of re_dfa_t's word_char. */
403
re_const_bitset_ptr_t word_char;
404
/* true if REG_ICASE. */
406
unsigned char is_utf8;
407
unsigned char map_notascii;
408
unsigned char mbs_allocated;
409
unsigned char offsets_needed;
410
unsigned char newline_anchor;
411
unsigned char word_ops_used;
414
typedef struct re_string_t re_string_t;
418
typedef struct re_dfa_t re_dfa_t;
421
# define IS_IN(libc) false
424
#define re_string_peek_byte(pstr, offset) \
425
((pstr)->mbs[(pstr)->cur_idx + offset])
426
#define re_string_fetch_byte(pstr) \
427
((pstr)->mbs[(pstr)->cur_idx++])
428
#define re_string_first_byte(pstr, idx) \
429
((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
430
#define re_string_is_single_byte_char(pstr, idx) \
431
((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
432
|| (pstr)->wcs[(idx) + 1] != WEOF))
433
#define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
434
#define re_string_cur_idx(pstr) ((pstr)->cur_idx)
435
#define re_string_get_buffer(pstr) ((pstr)->mbs)
436
#define re_string_length(pstr) ((pstr)->len)
437
#define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
438
#define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
439
#define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
442
# define MALLOC_0_IS_NONNULL 1
443
#elif !defined MALLOC_0_IS_NONNULL
444
# define MALLOC_0_IS_NONNULL 0
448
# define MAX(a,b) ((a) < (b) ? (b) : (a))
451
# define MIN(a,b) ((a) < (b) ? (a) : (b))
454
#define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
455
#define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
456
#define re_free(p) free (p)
460
struct bin_tree_t *parent;
461
struct bin_tree_t *left;
462
struct bin_tree_t *right;
463
struct bin_tree_t *first;
464
struct bin_tree_t *next;
468
/* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
469
Otherwise 'type' indicate the type of this node. */
472
typedef struct bin_tree_t bin_tree_t;
474
#define BIN_TREE_STORAGE_SIZE \
475
((1024 - sizeof (void *)) / sizeof (bin_tree_t))
477
struct bin_tree_storage_t
479
struct bin_tree_storage_t *next;
480
bin_tree_t data[BIN_TREE_STORAGE_SIZE];
482
typedef struct bin_tree_storage_t bin_tree_storage_t;
484
#define CONTEXT_WORD 1
485
#define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
486
#define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
487
#define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
489
#define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
490
#define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
491
#define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
492
#define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
493
#define IS_ORDINARY_CONTEXT(c) ((c) == 0)
495
#define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
496
#define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
497
#define IS_WIDE_WORD_CHAR(ch) (__iswalnum (ch) || (ch) == L'_')
498
#define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
500
#define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
501
((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
502
|| ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
503
|| ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
504
|| ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
506
#define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
507
((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
508
|| (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
509
|| (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
510
|| (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
516
re_node_set non_eps_nodes;
517
re_node_set inveclosure;
518
re_node_set *entrance_nodes;
519
struct re_dfastate_t **trtable, **word_trtable;
520
unsigned int context : 4;
521
unsigned int halt : 1;
522
/* If this state can accept "multi byte".
523
Note that we refer to multibyte characters, and multi character
524
collating elements as "multi byte". */
525
unsigned int accept_mb : 1;
526
/* If this state has backreference node(s). */
527
unsigned int has_backref : 1;
528
unsigned int has_constraint : 1;
530
typedef struct re_dfastate_t re_dfastate_t;
532
struct re_state_table_entry
536
re_dfastate_t **array;
539
/* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
545
re_dfastate_t **array;
548
/* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
553
Idx str_idx; /* The position NODE match at. */
555
} re_sub_match_last_t;
557
/* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
558
And information about the node, whose type is OP_CLOSE_SUBEXP,
559
corresponding to NODE is stored in LASTS. */
566
Idx alasts; /* Allocation size of LASTS. */
567
Idx nlasts; /* The number of LASTS. */
568
re_sub_match_last_t **lasts;
569
} re_sub_match_top_t;
571
struct re_backref_cache_entry
577
bitset_word_t eps_reachable_subexps_map;
583
/* The string object corresponding to the input string. */
585
const re_dfa_t *const dfa;
586
/* EFLAGS of the argument of regexec. */
588
/* Where the matching ends. */
591
/* The state log used by the matcher. */
592
re_dfastate_t **state_log;
594
/* Back reference cache. */
597
struct re_backref_cache_entry *bkref_ents;
601
re_sub_match_top_t **sub_tops;
602
} re_match_context_t;
606
re_dfastate_t **sifted_states;
607
re_dfastate_t **limited_states;
613
struct re_fail_stack_ent_t
618
re_node_set eps_via_nodes;
621
struct re_fail_stack_t
625
struct re_fail_stack_ent_t *stack;
636
re_node_set *eclosures;
637
re_node_set *inveclosures;
638
struct re_state_table_entry *state_table;
639
re_dfastate_t *init_state;
640
re_dfastate_t *init_state_word;
641
re_dfastate_t *init_state_nl;
642
re_dfastate_t *init_state_begbuf;
643
bin_tree_t *str_tree;
644
bin_tree_storage_t *str_tree_storage;
645
re_bitset_ptr_t sb_char;
646
int str_tree_storage_idx;
648
/* number of subexpressions 're_nsub' is in regex_t. */
649
re_hashval_t state_hash_mask;
651
Idx nbackref; /* The number of backreference in this dfa. */
653
/* Bitmap expressing which backreference is used. */
654
bitset_word_t used_bkref_map;
655
bitset_word_t completed_bkref_map;
657
unsigned int has_plural_match : 1;
658
/* If this dfa has "multibyte node", which is a backreference or
659
a node which can accept multibyte character or multi character
660
collating element. */
661
unsigned int has_mb_node : 1;
662
unsigned int is_utf8 : 1;
663
unsigned int map_notascii : 1;
664
unsigned int word_ops_used : 1;
675
#define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
676
#define re_node_set_remove(set,id) \
677
(re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
678
#define re_node_set_empty(p) ((p)->nelem = 0)
679
#define re_node_set_free(set) re_free ((set)->elems)
693
bracket_elem_type type;
703
/* Functions for bitset_t operation. */
706
bitset_set (bitset_t set, Idx i)
708
set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
712
bitset_clear (bitset_t set, Idx i)
714
set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
718
bitset_contain (const bitset_t set, Idx i)
720
return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
724
bitset_empty (bitset_t set)
726
memset (set, '\0', sizeof (bitset_t));
730
bitset_set_all (bitset_t set)
732
memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
733
if (SBC_MAX % BITSET_WORD_BITS != 0)
734
set[BITSET_WORDS - 1] =
735
((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
739
bitset_copy (bitset_t dest, const bitset_t src)
741
memcpy (dest, src, sizeof (bitset_t));
745
bitset_not (bitset_t set)
748
for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
749
set[bitset_i] = ~set[bitset_i];
750
if (SBC_MAX % BITSET_WORD_BITS != 0)
751
set[BITSET_WORDS - 1] =
752
((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
753
& ~set[BITSET_WORDS - 1]);
757
bitset_merge (bitset_t dest, const bitset_t src)
760
for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
761
dest[bitset_i] |= src[bitset_i];
765
bitset_mask (bitset_t dest, const bitset_t src)
768
for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
769
dest[bitset_i] &= src[bitset_i];
772
/* Functions for re_string. */
774
__attribute__ ((pure, unused))
775
re_string_char_size_at (const re_string_t *pstr, Idx idx)
778
if (pstr->mb_cur_max == 1)
780
for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
781
if (pstr->wcs[idx + byte_idx] != WEOF)
787
__attribute__ ((pure, unused))
788
re_string_wchar_at (const re_string_t *pstr, Idx idx)
790
if (pstr->mb_cur_max == 1)
791
return (wint_t) pstr->mbs[idx];
792
return (wint_t) pstr->wcs[idx];
796
# include <locale/weight.h>
800
__attribute__ ((pure, unused))
801
re_string_elem_size_at (const re_string_t *pstr, Idx idx)
804
const unsigned char *p, *extra;
805
const int32_t *table, *indirect;
806
uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
810
table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
811
extra = (const unsigned char *)
812
_NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
813
indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
814
_NL_COLLATE_INDIRECTMB);
816
findidx (table, indirect, extra, &p, pstr->len - idx);
817
return p - pstr->mbs - idx;
825
# if __glibc_has_attribute (__fallthrough__)
826
# define FALLTHROUGH __attribute__ ((__fallthrough__))
828
# define FALLTHROUGH ((void) 0)
831
# include "attribute.h"
834
#endif /* _REGEX_INTERNAL_H */
added
removed
support/regex_internal.h
