1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002-2011 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, write to the Free Software Foundation,
18 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
20 #ifndef _REGEX_INTERNAL_H
21 #define _REGEX_INTERNAL_H 1
32 # include "localcharset.h"
40 # include <bits/libc-lock.h>
42 # define __libc_lock_init(NAME) do { } while (0)
43 # define __libc_lock_lock(NAME) do { } while (0)
44 # define __libc_lock_unlock(NAME) do { } while (0)
47 /* In case that the system doesn't have isblank(). */
48 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
49 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
53 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
54 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
55 # include <locale/localeinfo.h>
56 # include <locale/elem-hash.h>
57 # include <locale/coll-lookup.h>
61 /* This is for other GNU distributions with internationalized messages. */
62 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
66 # define gettext(msgid) \
67 INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES)
70 # define gettext(msgid) (msgid)
74 /* This define is so xgettext can find the internationalizable
76 # define gettext_noop(String) String
79 /* For loser systems without the definition. */
81 # define SIZE_MAX ((size_t) -1)
84 #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE && HAVE_WCSCOLL) || _LIBC
85 # define RE_ENABLE_I18N
89 # define BE(expr, val) __builtin_expect (expr, val)
91 # define BE(expr, val) (expr)
97 /* Number of ASCII characters. */
98 #define ASCII_CHARS 0x80
100 /* Number of single byte characters. */
101 #define SBC_MAX (UCHAR_MAX + 1)
103 #define COLL_ELEM_LEN_MAX 8
105 /* The character which represents newline. */
106 #define NEWLINE_CHAR '\n'
107 #define WIDE_NEWLINE_CHAR L'\n'
109 /* Rename to standard API for using out of glibc. */
111 # define __wctype wctype
112 # define __iswctype iswctype
113 # define __btowc btowc
114 # define __wcrtomb wcrtomb
115 # define __mbrtowc mbrtowc
116 # define __regfree regfree
117 # define attribute_hidden
118 #endif /* not _LIBC */
120 #if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
121 # define __attribute(arg) __attribute__ (arg)
123 # define __attribute(arg)
126 typedef __re_idx_t Idx;
128 /* Special return value for failure to match. */
129 #define REG_MISSING ((Idx) -1)
131 /* Special return value for internal error. */
132 #define REG_ERROR ((Idx) -2)
134 /* Test whether N is a valid index, and is not one of the above. */
135 #ifdef _REGEX_LARGE_OFFSETS
136 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
138 # define REG_VALID_INDEX(n) (0 <= (n))
141 /* Test whether N is a valid nonzero index. */
142 #ifdef _REGEX_LARGE_OFFSETS
143 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
145 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
148 /* A hash value, suitable for computing hash tables. */
149 typedef __re_size_t re_hashval_t;
151 /* An integer used to represent a set of bits. It must be unsigned,
152 and must be at least as wide as unsigned int. */
153 typedef unsigned long int bitset_word_t;
154 /* All bits set in a bitset_word_t. */
155 #define BITSET_WORD_MAX ULONG_MAX
157 /* Number of bits in a bitset_word_t. For portability to hosts with
158 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
159 instead, deduce it directly from BITSET_WORD_MAX. Avoid
160 greater-than-32-bit integers and unconditional shifts by more than
161 31 bits, as they're not portable. */
162 #if BITSET_WORD_MAX == 0xffffffffUL
163 # define BITSET_WORD_BITS 32
164 #elif BITSET_WORD_MAX >> 31 >> 4 == 1
165 # define BITSET_WORD_BITS 36
166 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
167 # define BITSET_WORD_BITS 48
168 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
169 # define BITSET_WORD_BITS 60
170 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
171 # define BITSET_WORD_BITS 64
172 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
173 # define BITSET_WORD_BITS 72
174 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
175 # define BITSET_WORD_BITS 128
176 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
177 # define BITSET_WORD_BITS 256
178 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
179 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
180 # if BITSET_WORD_BITS <= SBC_MAX
181 # error "Invalid SBC_MAX"
184 # error "Add case for new bitset_word_t size"
187 /* Number of bitset_word_t values in a bitset_t. */
188 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
190 typedef bitset_word_t bitset_t[BITSET_WORDS];
191 typedef bitset_word_t *re_bitset_ptr_t;
192 typedef const bitset_word_t *re_const_bitset_ptr_t;
194 #define PREV_WORD_CONSTRAINT 0x0001
195 #define PREV_NOTWORD_CONSTRAINT 0x0002
196 #define NEXT_WORD_CONSTRAINT 0x0004
197 #define NEXT_NOTWORD_CONSTRAINT 0x0008
198 #define PREV_NEWLINE_CONSTRAINT 0x0010
199 #define NEXT_NEWLINE_CONSTRAINT 0x0020
200 #define PREV_BEGBUF_CONSTRAINT 0x0040
201 #define NEXT_ENDBUF_CONSTRAINT 0x0080
202 #define WORD_DELIM_CONSTRAINT 0x0100
203 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
207 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
208 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
209 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
210 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
211 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
212 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
213 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
214 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
215 WORD_DELIM = WORD_DELIM_CONSTRAINT,
216 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
230 /* Node type, These are used by token, node, tree. */
236 #ifdef RE_ENABLE_I18N
239 #endif /* RE_ENABLE_I18N */
241 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
242 when the debugger shows values of this enum type. */
243 #define EPSILON_BIT 8
244 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
245 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
246 OP_ALT = EPSILON_BIT | 2,
247 OP_DUP_ASTERISK = EPSILON_BIT | 3,
248 ANCHOR = EPSILON_BIT | 4,
250 /* Tree type, these are used only by tree. */
254 /* Token type, these are used only by token. */
266 OP_CLOSE_EQUIV_CLASS,
277 #ifdef RE_ENABLE_I18N
280 /* Multibyte characters. */
283 /* Collating symbols. */
288 /* Equivalence classes. */
290 int32_t *equiv_classes;
293 /* Range expressions. */
295 uint32_t *range_starts;
296 uint32_t *range_ends;
297 # else /* not _LIBC */
298 wchar_t *range_starts;
300 # endif /* not _LIBC */
302 /* Character classes. */
303 wctype_t *char_classes;
305 /* If this character set is the non-matching list. */
306 unsigned int non_match : 1;
308 /* # of multibyte characters. */
311 /* # of collating symbols. */
314 /* # of equivalence classes. */
317 /* # of range expressions. */
320 /* # of character classes. */
323 #endif /* RE_ENABLE_I18N */
329 unsigned char c; /* for CHARACTER */
330 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
331 #ifdef RE_ENABLE_I18N
332 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
333 #endif /* RE_ENABLE_I18N */
334 Idx idx; /* for BACK_REF */
335 re_context_type ctx_type; /* for ANCHOR */
337 #if __GNUC__ >= 2 && !__STRICT_ANSI__
338 re_token_type_t type : 8;
340 re_token_type_t type;
342 unsigned int constraint : 10; /* context constraint */
343 unsigned int duplicated : 1;
344 unsigned int opt_subexp : 1;
345 #ifdef RE_ENABLE_I18N
346 unsigned int accept_mb : 1;
347 /* These 2 bits can be moved into the union if needed (e.g. if running out
348 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
349 unsigned int mb_partial : 1;
351 unsigned int word_char : 1;
354 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
358 /* Indicate the raw buffer which is the original string passed as an
359 argument of regexec(), re_search(), etc.. */
360 const unsigned char *raw_mbs;
361 /* Store the multibyte string. In case of "case insensitive mode" like
362 REG_ICASE, upper cases of the string are stored, otherwise MBS points
363 the same address that RAW_MBS points. */
365 #ifdef RE_ENABLE_I18N
366 /* Store the wide character string which is corresponding to MBS. */
371 /* Index in RAW_MBS. Each character mbs[i] corresponds to
372 raw_mbs[raw_mbs_idx + i]. */
374 /* The length of the valid characters in the buffers. */
376 /* The corresponding number of bytes in raw_mbs array. */
378 /* The length of the buffers MBS and WCS. */
380 /* The index in MBS, which is updated by re_string_fetch_byte. */
382 /* length of RAW_MBS array. */
384 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
386 /* End of the buffer may be shorter than its length in the cases such
387 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
390 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
393 /* The context of mbs[0]. We store the context independently, since
394 the context of mbs[0] may be different from raw_mbs[0], which is
395 the beginning of the input string. */
396 unsigned int tip_context;
397 /* The translation passed as a part of an argument of re_compile_pattern. */
398 RE_TRANSLATE_TYPE trans;
399 /* Copy of re_dfa_t's word_char. */
400 re_const_bitset_ptr_t word_char;
401 /* true if REG_ICASE. */
403 unsigned char is_utf8;
404 unsigned char map_notascii;
405 unsigned char mbs_allocated;
406 unsigned char offsets_needed;
407 unsigned char newline_anchor;
408 unsigned char word_ops_used;
411 typedef struct re_string_t re_string_t;
415 typedef struct re_dfa_t re_dfa_t;
418 # if defined __i386__ && !defined __EMX__
419 # define internal_function __attribute ((regparm (3), stdcall))
421 # define internal_function
425 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
428 #ifdef RE_ENABLE_I18N
429 static void build_wcs_buffer (re_string_t *pstr) internal_function;
430 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
432 #endif /* RE_ENABLE_I18N */
433 static void build_upper_buffer (re_string_t *pstr) internal_function;
434 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
435 static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
437 internal_function __attribute ((pure));
438 #define re_string_peek_byte(pstr, offset) \
439 ((pstr)->mbs[(pstr)->cur_idx + offset])
440 #define re_string_fetch_byte(pstr) \
441 ((pstr)->mbs[(pstr)->cur_idx++])
442 #define re_string_first_byte(pstr, idx) \
443 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
444 #define re_string_is_single_byte_char(pstr, idx) \
445 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
446 || (pstr)->wcs[(idx) + 1] != WEOF))
447 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
448 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
449 #define re_string_get_buffer(pstr) ((pstr)->mbs)
450 #define re_string_length(pstr) ((pstr)->len)
451 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
452 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
453 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
459 /* The OS usually guarantees only one guard page at the bottom of the stack,
460 and a page size can be as small as 4096 bytes. So we cannot safely
461 allocate anything larger than 4096 bytes. Also care for the possibility
462 of a few compiler-allocated temporary stack slots. */
463 # define __libc_use_alloca(n) ((n) < 4032)
465 /* alloca is implemented with malloc, so just use malloc. */
466 # define __libc_use_alloca(n) 0
468 # define alloca(n) malloc (n)
473 # define MAX(a,b) ((a) < (b) ? (b) : (a))
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;
700 __libc_lock_define (, lock)
704 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
705 #define re_node_set_remove(set,id) \
706 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
707 #define re_node_set_empty(p) ((p)->nelem = 0)
708 #define re_node_set_free(set) re_free ((set)->elems)
722 bracket_elem_type type;
732 /* Inline functions for bitset_t operation. */
735 bitset_set (bitset_t set, Idx i)
737 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
741 bitset_clear (bitset_t set, Idx i)
743 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
747 bitset_contain (const bitset_t set, Idx i)
749 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
753 bitset_empty (bitset_t set)
755 memset (set, '\0', sizeof (bitset_t));
759 bitset_set_all (bitset_t set)
761 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
762 if (SBC_MAX % BITSET_WORD_BITS != 0)
763 set[BITSET_WORDS - 1] =
764 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
768 bitset_copy (bitset_t dest, const bitset_t src)
770 memcpy (dest, src, sizeof (bitset_t));
774 bitset_not (bitset_t set)
777 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
778 set[bitset_i] = ~set[bitset_i];
779 if (SBC_MAX % BITSET_WORD_BITS != 0)
780 set[BITSET_WORDS - 1] =
781 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
782 & ~set[BITSET_WORDS - 1]);
786 bitset_merge (bitset_t dest, const bitset_t src)
789 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
790 dest[bitset_i] |= src[bitset_i];
794 bitset_mask (bitset_t dest, const bitset_t src)
797 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
798 dest[bitset_i] &= src[bitset_i];
801 #ifdef RE_ENABLE_I18N
802 /* Inline functions for re_string. */
804 internal_function __attribute ((pure))
805 re_string_char_size_at (const re_string_t *pstr, Idx idx)
808 if (pstr->mb_cur_max == 1)
810 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
811 if (pstr->wcs[idx + byte_idx] != WEOF)
817 internal_function __attribute ((pure))
818 re_string_wchar_at (const re_string_t *pstr, Idx idx)
820 if (pstr->mb_cur_max == 1)
821 return (wint_t) pstr->mbs[idx];
822 return (wint_t) pstr->wcs[idx];
826 internal_function __attribute ((pure))
827 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
830 const unsigned char *p, *extra;
831 const int32_t *table, *indirect;
833 # include <locale/weight.h>
834 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
838 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
839 extra = (const unsigned char *)
840 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
841 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
842 _NL_COLLATE_INDIRECTMB);
845 return p - pstr->mbs - idx;
851 #endif /* RE_ENABLE_I18N */
853 #ifndef __GNUC_PREREQ
854 # if defined __GNUC__ && defined __GNUC_MINOR__
855 # define __GNUC_PREREQ(maj, min) \
856 ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
858 # define __GNUC_PREREQ(maj, min) 0
862 #if __GNUC_PREREQ (3,4)
863 # undef __attribute_warn_unused_result__
864 # define __attribute_warn_unused_result__ \
865 __attribute__ ((__warn_unused_result__))
867 # define __attribute_warn_unused_result__ /* empty */
870 #endif /* _REGEX_INTERNAL_H */