1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002-2012 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4 Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2, or (at your option)
11 This program 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
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License along
17 with this program; if not, see <http://www.gnu.org/licenses/>. */
19 #ifndef _REGEX_INTERNAL_H
20 #define _REGEX_INTERNAL_H 1
30 # include "localcharset.h"
38 # include <bits/libc-lock.h>
40 # define __libc_lock_define(CLASS,NAME)
41 # define __libc_lock_init(NAME) do { } while (0)
42 # define __libc_lock_lock(NAME) do { } while (0)
43 # define __libc_lock_unlock(NAME) do { } while (0)
46 /* In case that the system doesn't have isblank(). */
47 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
48 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
52 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
53 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
54 # include <locale/localeinfo.h>
55 # include <locale/elem-hash.h>
56 # include <locale/coll-lookup.h>
60 /* This is for other GNU distributions with internationalized messages. */
61 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
65 # define gettext(msgid) \
66 INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES)
69 # define gettext(msgid) (msgid)
73 /* This define is so xgettext can find the internationalizable
75 # define gettext_noop(String) String
78 #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE && HAVE_WCSCOLL) || _LIBC
79 # define RE_ENABLE_I18N
83 # define BE(expr, val) __builtin_expect (expr, val)
85 # define BE(expr, val) (expr)
91 /* Number of ASCII characters. */
92 #define ASCII_CHARS 0x80
94 /* Number of single byte characters. */
95 #define SBC_MAX (UCHAR_MAX + 1)
97 #define COLL_ELEM_LEN_MAX 8
99 /* The character which represents newline. */
100 #define NEWLINE_CHAR '\n'
101 #define WIDE_NEWLINE_CHAR L'\n'
103 /* Rename to standard API for using out of glibc. */
105 # define __wctype wctype
106 # define __iswctype iswctype
107 # define __btowc btowc
108 # define __mbrtowc mbrtowc
109 # define __wcrtomb wcrtomb
110 # define __regfree regfree
111 # define attribute_hidden
112 #endif /* not _LIBC */
114 #if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
115 # define __attribute(arg) __attribute__ (arg)
117 # define __attribute(arg)
120 typedef __re_idx_t Idx;
121 #ifdef _REGEX_LARGE_OFFSETS
122 # define IDX_MAX (SIZE_MAX - 2)
124 # define IDX_MAX INT_MAX
127 /* Special return value for failure to match. */
128 #define REG_MISSING ((Idx) -1)
130 /* Special return value for internal error. */
131 #define REG_ERROR ((Idx) -2)
133 /* Test whether N is a valid index, and is not one of the above. */
134 #ifdef _REGEX_LARGE_OFFSETS
135 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
137 # define REG_VALID_INDEX(n) (0 <= (n))
140 /* Test whether N is a valid nonzero index. */
141 #ifdef _REGEX_LARGE_OFFSETS
142 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
144 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
147 /* A hash value, suitable for computing hash tables. */
148 typedef __re_size_t re_hashval_t;
150 /* An integer used to represent a set of bits. It must be unsigned,
151 and must be at least as wide as unsigned int. */
152 typedef unsigned long int bitset_word_t;
153 /* All bits set in a bitset_word_t. */
154 #define BITSET_WORD_MAX ULONG_MAX
156 /* Number of bits in a bitset_word_t. For portability to hosts with
157 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
158 instead, deduce it directly from BITSET_WORD_MAX. Avoid
159 greater-than-32-bit integers and unconditional shifts by more than
160 31 bits, as they're not portable. */
161 #if BITSET_WORD_MAX == 0xffffffffUL
162 # define BITSET_WORD_BITS 32
163 #elif BITSET_WORD_MAX >> 31 >> 4 == 1
164 # define BITSET_WORD_BITS 36
165 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
166 # define BITSET_WORD_BITS 48
167 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
168 # define BITSET_WORD_BITS 60
169 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
170 # define BITSET_WORD_BITS 64
171 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
172 # define BITSET_WORD_BITS 72
173 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
174 # define BITSET_WORD_BITS 128
175 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
176 # define BITSET_WORD_BITS 256
177 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
178 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
179 # if BITSET_WORD_BITS <= SBC_MAX
180 # error "Invalid SBC_MAX"
183 # error "Add case for new bitset_word_t size"
186 /* Number of bitset_word_t values in a bitset_t. */
187 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
189 typedef bitset_word_t bitset_t[BITSET_WORDS];
190 typedef bitset_word_t *re_bitset_ptr_t;
191 typedef const bitset_word_t *re_const_bitset_ptr_t;
193 #define PREV_WORD_CONSTRAINT 0x0001
194 #define PREV_NOTWORD_CONSTRAINT 0x0002
195 #define NEXT_WORD_CONSTRAINT 0x0004
196 #define NEXT_NOTWORD_CONSTRAINT 0x0008
197 #define PREV_NEWLINE_CONSTRAINT 0x0010
198 #define NEXT_NEWLINE_CONSTRAINT 0x0020
199 #define PREV_BEGBUF_CONSTRAINT 0x0040
200 #define NEXT_ENDBUF_CONSTRAINT 0x0080
201 #define WORD_DELIM_CONSTRAINT 0x0100
202 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
206 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
207 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
208 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
209 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
210 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
211 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
212 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
213 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
214 WORD_DELIM = WORD_DELIM_CONSTRAINT,
215 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
229 /* Node type, These are used by token, node, tree. */
235 #ifdef RE_ENABLE_I18N
238 #endif /* RE_ENABLE_I18N */
240 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
241 when the debugger shows values of this enum type. */
242 #define EPSILON_BIT 8
243 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
244 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
245 OP_ALT = EPSILON_BIT | 2,
246 OP_DUP_ASTERISK = EPSILON_BIT | 3,
247 ANCHOR = EPSILON_BIT | 4,
249 /* Tree type, these are used only by tree. */
253 /* Token type, these are used only by token. */
265 OP_CLOSE_EQUIV_CLASS,
276 #ifdef RE_ENABLE_I18N
279 /* Multibyte characters. */
282 /* Collating symbols. */
287 /* Equivalence classes. */
289 int32_t *equiv_classes;
292 /* Range expressions. */
294 uint32_t *range_starts;
295 uint32_t *range_ends;
296 # else /* not _LIBC */
297 wchar_t *range_starts;
299 # endif /* not _LIBC */
301 /* Character classes. */
302 wctype_t *char_classes;
304 /* If this character set is the non-matching list. */
305 unsigned int non_match : 1;
307 /* # of multibyte characters. */
310 /* # of collating symbols. */
313 /* # of equivalence classes. */
316 /* # of range expressions. */
319 /* # of character classes. */
322 #endif /* RE_ENABLE_I18N */
328 unsigned char c; /* for CHARACTER */
329 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
330 #ifdef RE_ENABLE_I18N
331 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
332 #endif /* RE_ENABLE_I18N */
333 Idx idx; /* for BACK_REF */
334 re_context_type ctx_type; /* for ANCHOR */
336 #if __GNUC__ >= 2 && !defined __STRICT_ANSI__
337 re_token_type_t type : 8;
339 re_token_type_t type;
341 unsigned int constraint : 10; /* context constraint */
342 unsigned int duplicated : 1;
343 unsigned int opt_subexp : 1;
344 #ifdef RE_ENABLE_I18N
345 unsigned int accept_mb : 1;
346 /* These 2 bits can be moved into the union if needed (e.g. if running out
347 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
348 unsigned int mb_partial : 1;
350 unsigned int word_char : 1;
353 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
357 /* Indicate the raw buffer which is the original string passed as an
358 argument of regexec(), re_search(), etc.. */
359 const unsigned char *raw_mbs;
360 /* Store the multibyte string. In case of "case insensitive mode" like
361 REG_ICASE, upper cases of the string are stored, otherwise MBS points
362 the same address that RAW_MBS points. */
364 #ifdef RE_ENABLE_I18N
365 /* Store the wide character string which is corresponding to MBS. */
370 /* Index in RAW_MBS. Each character mbs[i] corresponds to
371 raw_mbs[raw_mbs_idx + i]. */
373 /* The length of the valid characters in the buffers. */
375 /* The corresponding number of bytes in raw_mbs array. */
377 /* The length of the buffers MBS and WCS. */
379 /* The index in MBS, which is updated by re_string_fetch_byte. */
381 /* length of RAW_MBS array. */
383 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
385 /* End of the buffer may be shorter than its length in the cases such
386 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
389 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
392 /* The context of mbs[0]. We store the context independently, since
393 the context of mbs[0] may be different from raw_mbs[0], which is
394 the beginning of the input string. */
395 unsigned int tip_context;
396 /* The translation passed as a part of an argument of re_compile_pattern. */
397 RE_TRANSLATE_TYPE trans;
398 /* Copy of re_dfa_t's word_char. */
399 re_const_bitset_ptr_t word_char;
400 /* true if REG_ICASE. */
402 unsigned char is_utf8;
403 unsigned char map_notascii;
404 unsigned char mbs_allocated;
405 unsigned char offsets_needed;
406 unsigned char newline_anchor;
407 unsigned char word_ops_used;
410 typedef struct re_string_t re_string_t;
414 typedef struct re_dfa_t re_dfa_t;
417 # define internal_function
421 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
424 # ifdef RE_ENABLE_I18N
425 static void build_wcs_buffer (re_string_t *pstr) internal_function;
426 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
428 # endif /* RE_ENABLE_I18N */
429 static void build_upper_buffer (re_string_t *pstr) internal_function;
430 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
431 static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
433 internal_function __attribute ((pure));
435 #define re_string_peek_byte(pstr, offset) \
436 ((pstr)->mbs[(pstr)->cur_idx + offset])
437 #define re_string_fetch_byte(pstr) \
438 ((pstr)->mbs[(pstr)->cur_idx++])
439 #define re_string_first_byte(pstr, idx) \
440 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
441 #define re_string_is_single_byte_char(pstr, idx) \
442 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
443 || (pstr)->wcs[(idx) + 1] != WEOF))
444 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
445 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
446 #define re_string_get_buffer(pstr) ((pstr)->mbs)
447 #define re_string_length(pstr) ((pstr)->len)
448 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
449 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
450 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
456 /* The OS usually guarantees only one guard page at the bottom of the stack,
457 and a page size can be as small as 4096 bytes. So we cannot safely
458 allocate anything larger than 4096 bytes. Also care for the possibility
459 of a few compiler-allocated temporary stack slots. */
460 # define __libc_use_alloca(n) ((n) < 4032)
462 /* alloca is implemented with malloc, so just use malloc. */
463 # define __libc_use_alloca(n) 0
465 # define alloca(n) malloc (n)
470 # define MAX(a,b) ((a) < (b) ? (b) : (a))
473 # define MIN(a,b) ((a) < (b) ? (a) : (b))
476 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
477 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
478 #define re_free(p) free (p)
482 struct bin_tree_t *parent;
483 struct bin_tree_t *left;
484 struct bin_tree_t *right;
485 struct bin_tree_t *first;
486 struct bin_tree_t *next;
490 /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
491 Otherwise 'type' indicate the type of this node. */
494 typedef struct bin_tree_t bin_tree_t;
496 #define BIN_TREE_STORAGE_SIZE \
497 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
499 struct bin_tree_storage_t
501 struct bin_tree_storage_t *next;
502 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
504 typedef struct bin_tree_storage_t bin_tree_storage_t;
506 #define CONTEXT_WORD 1
507 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
508 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
509 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
511 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
512 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
513 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
514 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
515 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
517 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
518 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
519 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
520 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
522 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
523 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
524 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
525 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
526 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
528 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
529 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
530 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
531 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
532 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
538 re_node_set non_eps_nodes;
539 re_node_set inveclosure;
540 re_node_set *entrance_nodes;
541 struct re_dfastate_t **trtable, **word_trtable;
542 unsigned int context : 4;
543 unsigned int halt : 1;
544 /* If this state can accept "multi byte".
545 Note that we refer to multibyte characters, and multi character
546 collating elements as "multi byte". */
547 unsigned int accept_mb : 1;
548 /* If this state has backreference node(s). */
549 unsigned int has_backref : 1;
550 unsigned int has_constraint : 1;
552 typedef struct re_dfastate_t re_dfastate_t;
554 struct re_state_table_entry
558 re_dfastate_t **array;
561 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
567 re_dfastate_t **array;
570 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
575 Idx str_idx; /* The position NODE match at. */
577 } re_sub_match_last_t;
579 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
580 And information about the node, whose type is OP_CLOSE_SUBEXP,
581 corresponding to NODE is stored in LASTS. */
588 Idx alasts; /* Allocation size of LASTS. */
589 Idx nlasts; /* The number of LASTS. */
590 re_sub_match_last_t **lasts;
591 } re_sub_match_top_t;
593 struct re_backref_cache_entry
601 unsigned short int eps_reachable_subexps_map;
606 /* The string object corresponding to the input string. */
608 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
609 const re_dfa_t *const dfa;
613 /* EFLAGS of the argument of regexec. */
615 /* Where the matching ends. */
618 /* The state log used by the matcher. */
619 re_dfastate_t **state_log;
621 /* Back reference cache. */
624 struct re_backref_cache_entry *bkref_ents;
628 re_sub_match_top_t **sub_tops;
629 } re_match_context_t;
633 re_dfastate_t **sifted_states;
634 re_dfastate_t **limited_states;
640 struct re_fail_stack_ent_t
645 re_node_set eps_via_nodes;
648 struct re_fail_stack_t
652 struct re_fail_stack_ent_t *stack;
663 re_node_set *eclosures;
664 re_node_set *inveclosures;
665 struct re_state_table_entry *state_table;
666 re_dfastate_t *init_state;
667 re_dfastate_t *init_state_word;
668 re_dfastate_t *init_state_nl;
669 re_dfastate_t *init_state_begbuf;
670 bin_tree_t *str_tree;
671 bin_tree_storage_t *str_tree_storage;
672 re_bitset_ptr_t sb_char;
673 int str_tree_storage_idx;
675 /* number of subexpressions 're_nsub' is in regex_t. */
676 re_hashval_t state_hash_mask;
678 Idx nbackref; /* The number of backreference in this dfa. */
680 /* Bitmap expressing which backreference is used. */
681 bitset_word_t used_bkref_map;
682 bitset_word_t completed_bkref_map;
684 unsigned int has_plural_match : 1;
685 /* If this dfa has "multibyte node", which is a backreference or
686 a node which can accept multibyte character or multi character
687 collating element. */
688 unsigned int has_mb_node : 1;
689 unsigned int is_utf8 : 1;
690 unsigned int map_notascii : 1;
691 unsigned int word_ops_used : 1;
699 __libc_lock_define (, lock)
702 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
703 #define re_node_set_remove(set,id) \
704 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
705 #define re_node_set_empty(p) ((p)->nelem = 0)
706 #define re_node_set_free(set) re_free ((set)->elems)
720 bracket_elem_type type;
730 /* Inline functions for bitset_t operation. */
733 bitset_set (bitset_t set, Idx i)
735 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
739 bitset_clear (bitset_t set, Idx i)
741 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
745 bitset_contain (const bitset_t set, Idx i)
747 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
751 bitset_empty (bitset_t set)
753 memset (set, '\0', sizeof (bitset_t));
757 bitset_set_all (bitset_t set)
759 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
760 if (SBC_MAX % BITSET_WORD_BITS != 0)
761 set[BITSET_WORDS - 1] =
762 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
766 bitset_copy (bitset_t dest, const bitset_t src)
768 memcpy (dest, src, sizeof (bitset_t));
772 bitset_not (bitset_t set)
775 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
776 set[bitset_i] = ~set[bitset_i];
777 if (SBC_MAX % BITSET_WORD_BITS != 0)
778 set[BITSET_WORDS - 1] =
779 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
780 & ~set[BITSET_WORDS - 1]);
784 bitset_merge (bitset_t dest, const bitset_t src)
787 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
788 dest[bitset_i] |= src[bitset_i];
792 bitset_mask (bitset_t dest, const bitset_t src)
795 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
796 dest[bitset_i] &= src[bitset_i];
799 #ifdef RE_ENABLE_I18N
800 /* Inline functions for re_string. */
802 internal_function __attribute ((pure))
803 re_string_char_size_at (const re_string_t *pstr, Idx idx)
806 if (pstr->mb_cur_max == 1)
808 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
809 if (pstr->wcs[idx + byte_idx] != WEOF)
815 internal_function __attribute ((pure))
816 re_string_wchar_at (const re_string_t *pstr, Idx idx)
818 if (pstr->mb_cur_max == 1)
819 return (wint_t) pstr->mbs[idx];
820 return (wint_t) pstr->wcs[idx];
825 internal_function __attribute ((pure))
826 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
829 const unsigned char *p, *extra;
830 const int32_t *table, *indirect;
831 # include <locale/weight.h>
832 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
836 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
837 extra = (const unsigned char *)
838 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
839 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
840 _NL_COLLATE_INDIRECTMB);
842 findidx (&p, pstr->len - idx);
843 return p - pstr->mbs - idx;
850 #endif /* RE_ENABLE_I18N */
852 #ifndef __GNUC_PREREQ
853 # if defined __GNUC__ && defined __GNUC_MINOR__
854 # define __GNUC_PREREQ(maj, min) \
855 ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
857 # define __GNUC_PREREQ(maj, min) 0
861 #if __GNUC_PREREQ (3,4)
862 # undef __attribute_warn_unused_result__
863 # define __attribute_warn_unused_result__ \
864 __attribute__ ((__warn_unused_result__))
866 # define __attribute_warn_unused_result__ /* empty */
869 #endif /* _REGEX_INTERNAL_H */