1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002, 2003, 2004, 2005 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
34 #if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET || defined _LIBC
35 # include <langinfo.h>
37 #if defined HAVE_LOCALE_H || defined _LIBC
40 #if defined HAVE_WCHAR_H || defined _LIBC
42 #endif /* HAVE_WCHAR_H || _LIBC */
43 #if defined HAVE_WCTYPE_H || defined _LIBC
45 #endif /* HAVE_WCTYPE_H || _LIBC */
47 # include <bits/libc-lock.h>
49 # define __libc_lock_define(CLASS,NAME)
50 # define __libc_lock_init(NAME) do { } while (0)
51 # define __libc_lock_lock(NAME) do { } while (0)
52 # define __libc_lock_unlock(NAME) do { } while (0)
55 /* In case that the system doesn't have isblank(). */
56 #if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank
57 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
61 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
62 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
63 # include <locale/localeinfo.h>
64 # include <locale/elem-hash.h>
65 # include <locale/coll-lookup.h>
69 /* This is for other GNU distributions with internationalized messages. */
70 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
74 # define gettext(msgid) \
75 INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES)
78 # define gettext(msgid) (msgid)
82 /* This define is so xgettext can find the internationalizable
84 # define gettext_noop(String) String
87 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC
88 # define RE_ENABLE_I18N
92 # if defined __GNUC__ && defined __GNUC_MINOR__
93 # define __GNUC_PREREQ(maj, min) \
94 ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
96 # define __GNUC_PREREQ(maj, min) 0
100 #if !__GNUC_PREREQ (3, 1)
101 # define always_inline
105 # define BE(expr, val) __builtin_expect (expr, val)
107 # define BE(expr, val) (expr)
112 /* Number of single byte character. */
115 #define COLL_ELEM_LEN_MAX 8
117 /* The character which represents newline. */
118 #define NEWLINE_CHAR '\n'
119 #define WIDE_NEWLINE_CHAR L'\n'
121 /* Rename to standard API for using out of glibc. */
123 # define __wctype wctype
124 # define __iswctype iswctype
125 # define __btowc btowc
127 # define __mempcpy mempcpy
129 # define __wcrtomb wcrtomb
130 # define __regfree regfree
131 # define attribute_hidden
132 #endif /* not _LIBC */
135 # define __attribute(arg) __attribute__ (arg)
137 # define __attribute(arg)
140 extern const char __re_error_msgid[] attribute_hidden;
141 extern const size_t __re_error_msgid_idx[] attribute_hidden;
143 typedef __re_idx_t Idx;
145 /* Special return value for failure to match. */
146 #define REG_MISSING ((Idx) -1)
148 /* Special return value for internal error. */
149 #define REG_ERROR ((Idx) -2)
151 /* Test whether N is a valid index, and is not one of the above. */
152 #ifdef _REGEX_LARGE_OFFSETS
153 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
155 # define REG_VALID_INDEX(n) (0 <= (n))
158 /* Test whether N is a valid nonzero index. */
159 #ifdef _REGEX_LARGE_OFFSETS
160 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
162 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
165 /* A hash value, suitable for computing hash tables. */
166 typedef __re_size_t re_hashval_t;
168 /* An integer used to represent a set of bits. It must be unsigned,
169 and must be at least as wide as unsigned int. */
170 typedef unsigned long int bitset_word;
172 /* Maximum value of a bitset word. It must be useful in preprocessor
173 contexts, and must be consistent with bitset_word. */
174 #define BITSET_WORD_MAX ULONG_MAX
176 /* Number of bits in a bitset word. Avoid greater-than-32-bit
177 integers and unconditional shifts by more than 31 bits, as they're
179 #if BITSET_WORD_MAX == 0xffffffff
180 # define BITSET_WORD_BITS 32
181 #elif BITSET_WORD_MAX >> 31 >> 5 == 1
182 # define BITSET_WORD_BITS 36
183 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
184 # define BITSET_WORD_BITS 48
185 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
186 # define BITSET_WORD_BITS 60
187 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
188 # define BITSET_WORD_BITS 64
189 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
190 # define BITSET_WORD_BITS 72
191 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
192 # define BITSET_WORD_BITS 128
193 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
194 # define BITSET_WORD_BITS 256
195 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
196 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
197 # if BITSET_WORD_BITS <= SBC_MAX
198 # error "Invalid SBC_MAX"
201 # error "Add case for new bitset_word size"
204 /* Number of bitset words in a bitset. */
205 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
207 typedef bitset_word bitset[BITSET_WORDS];
208 typedef bitset_word *re_bitset_ptr_t;
209 typedef const bitset_word *re_const_bitset_ptr_t;
211 #define PREV_WORD_CONSTRAINT 0x0001
212 #define PREV_NOTWORD_CONSTRAINT 0x0002
213 #define NEXT_WORD_CONSTRAINT 0x0004
214 #define NEXT_NOTWORD_CONSTRAINT 0x0008
215 #define PREV_NEWLINE_CONSTRAINT 0x0010
216 #define NEXT_NEWLINE_CONSTRAINT 0x0020
217 #define PREV_BEGBUF_CONSTRAINT 0x0040
218 #define NEXT_ENDBUF_CONSTRAINT 0x0080
219 #define WORD_DELIM_CONSTRAINT 0x0100
220 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
224 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
225 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
226 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
227 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
228 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
229 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
230 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
231 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
232 WORD_DELIM = WORD_DELIM_CONSTRAINT,
233 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
247 /* Node type, These are used by token, node, tree. */
253 #ifdef RE_ENABLE_I18N
256 #endif /* RE_ENABLE_I18N */
258 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
259 when the debugger shows values of this enum type. */
260 #define EPSILON_BIT 8
261 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
262 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
263 OP_ALT = EPSILON_BIT | 2,
264 OP_DUP_ASTERISK = EPSILON_BIT | 3,
265 ANCHOR = EPSILON_BIT | 4,
267 /* Tree type, these are used only by tree. */
271 /* Token type, these are used only by token. */
283 OP_CLOSE_EQUIV_CLASS,
294 #ifdef RE_ENABLE_I18N
297 /* Multibyte characters. */
300 /* Collating symbols. */
305 /* Equivalence classes. */
307 int32_t *equiv_classes;
310 /* Range expressions. */
312 uint32_t *range_starts;
313 uint32_t *range_ends;
314 # else /* not _LIBC */
315 wchar_t *range_starts;
317 # endif /* not _LIBC */
319 /* Character classes. */
320 wctype_t *char_classes;
322 /* If this character set is the non-matching list. */
323 unsigned int non_match : 1;
325 /* # of multibyte characters. */
328 /* # of collating symbols. */
331 /* # of equivalence classes. */
334 /* # of range expressions. */
337 /* # of character classes. */
340 #endif /* RE_ENABLE_I18N */
346 unsigned char c; /* for CHARACTER */
347 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
348 #ifdef RE_ENABLE_I18N
349 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
350 #endif /* RE_ENABLE_I18N */
351 Idx idx; /* for BACK_REF */
352 re_context_type ctx_type; /* for ANCHOR */
355 re_token_type_t type : 8;
357 re_token_type_t type;
359 unsigned int constraint : 10; /* context constraint */
360 unsigned int duplicated : 1;
361 unsigned int opt_subexp : 1;
362 #ifdef RE_ENABLE_I18N
363 unsigned int accept_mb : 1;
364 /* These 2 bits can be moved into the union if needed (e.g. if running out
365 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
366 unsigned int mb_partial : 1;
368 unsigned int word_char : 1;
371 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
375 /* Indicate the raw buffer which is the original string passed as an
376 argument of regexec(), re_search(), etc.. */
377 const unsigned char *raw_mbs;
378 /* Store the multibyte string. In case of "case insensitive mode" like
379 REG_ICASE, upper cases of the string are stored, otherwise MBS points
380 the same address that RAW_MBS points. */
382 #ifdef RE_ENABLE_I18N
383 /* Store the wide character string which is corresponding to MBS. */
388 /* Index in RAW_MBS. Each character mbs[i] corresponds to
389 raw_mbs[raw_mbs_idx + i]. */
391 /* The length of the valid characters in the buffers. */
393 /* The corresponding number of bytes in raw_mbs array. */
395 /* The length of the buffers MBS and WCS. */
397 /* The index in MBS, which is updated by re_string_fetch_byte. */
399 /* length of RAW_MBS array. */
401 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
403 /* End of the buffer may be shorter than its length in the cases such
404 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
407 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
410 /* The context of mbs[0]. We store the context independently, since
411 the context of mbs[0] may be different from raw_mbs[0], which is
412 the beginning of the input string. */
413 unsigned int tip_context;
414 /* The translation passed as a part of an argument of re_compile_pattern. */
415 unsigned REG_TRANSLATE_TYPE trans;
416 /* Copy of re_dfa_t's word_char. */
417 re_const_bitset_ptr_t word_char;
418 /* true if REG_ICASE. */
420 unsigned char is_utf8;
421 unsigned char map_notascii;
422 unsigned char mbs_allocated;
423 unsigned char offsets_needed;
424 unsigned char newline_anchor;
425 unsigned char word_ops_used;
428 typedef struct re_string_t re_string_t;
432 typedef struct re_dfa_t re_dfa_t;
436 # define internal_function __attribute ((regparm (3), stdcall))
438 # define internal_function
442 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
445 #ifdef RE_ENABLE_I18N
446 static void build_wcs_buffer (re_string_t *pstr) internal_function;
447 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
449 #endif /* RE_ENABLE_I18N */
450 static void build_upper_buffer (re_string_t *pstr) internal_function;
451 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
452 static unsigned int re_string_context_at (const re_string_t *input,
454 internal_function __attribute ((pure));
456 #define re_string_peek_byte(pstr, offset) \
457 ((pstr)->mbs[(pstr)->cur_idx + offset])
458 #define re_string_fetch_byte(pstr) \
459 ((pstr)->mbs[(pstr)->cur_idx++])
460 #define re_string_first_byte(pstr, idx) \
461 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
462 #define re_string_is_single_byte_char(pstr, idx) \
463 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
464 || (pstr)->wcs[(idx) + 1] != WEOF))
465 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
466 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
467 #define re_string_get_buffer(pstr) ((pstr)->mbs)
468 #define re_string_length(pstr) ((pstr)->len)
469 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
470 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
471 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
477 /* The OS usually guarantees only one guard page at the bottom of the stack,
478 and a page size can be as small as 4096 bytes. So we cannot safely
479 allocate anything larger than 4096 bytes. Also care for the possibility
480 of a few compiler-allocated temporary stack slots. */
481 # define __libc_use_alloca(n) ((n) < 4032)
483 /* alloca is implemented with malloc, so just use malloc. */
484 # define __libc_use_alloca(n) 0
488 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
489 #define re_xmalloc(t,n) ((t *) re_xnmalloc (n, sizeof (t)))
490 #define re_calloc(t,n) ((t *) calloc (n, sizeof (t)))
491 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
492 #define re_xrealloc(p,t,n) ((t *) re_xnrealloc (p, n, sizeof (t)))
493 #define re_x2realloc(p,t,pn) ((t *) re_x2nrealloc (p, pn, sizeof (t)))
494 #define re_free(p) free (p)
497 # define SIZE_MAX ((size_t) -1)
500 /* Return true if an array of N objects, each of size S, cannot exist
501 due to size arithmetic overflow. S must be nonzero. */
503 re_alloc_oversized (size_t n, size_t s)
505 return BE (SIZE_MAX / s < n, 0);
508 /* Return true if an array of (2 * N + 1) objects, each of size S,
509 cannot exist due to size arithmetic overflow. S must be nonzero. */
511 re_x2alloc_oversized (size_t n, size_t s)
513 return BE ((SIZE_MAX / s - 1) / 2 < n, 0);
516 /* Allocate an array of N objects, each with S bytes of memory,
517 dynamically, with error checking. S must be nonzero. */
519 re_xnmalloc (size_t n, size_t s)
521 return re_alloc_oversized (n, s) ? NULL : malloc (n * s);
524 /* Change the size of an allocated block of memory P to an array of N
525 objects each of S bytes, with error checking. S must be nonzero. */
527 re_xnrealloc (void *p, size_t n, size_t s)
529 return re_alloc_oversized (n, s) ? NULL : realloc (p, n * s);
532 /* Reallocate a block of memory P to an array of (2 * (*PN) + 1)
533 objects each of S bytes, with error checking. S must be nonzero.
534 If the allocation is successful, set *PN to the new allocation
535 count and return the resulting pointer. Otherwise, return
538 re_x2nrealloc (void *p, size_t *pn, size_t s)
540 if (re_x2alloc_oversized (*pn, s))
544 /* Add 1 in case *PN is zero. */
545 size_t n1 = 2 * *pn + 1;
546 p = realloc (p, n1 * s);
547 if (BE (p != NULL, 1))
555 struct bin_tree_t *parent;
556 struct bin_tree_t *left;
557 struct bin_tree_t *right;
558 struct bin_tree_t *first;
559 struct bin_tree_t *next;
563 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
564 Otherwise `type' indicate the type of this node. */
567 typedef struct bin_tree_t bin_tree_t;
569 #define BIN_TREE_STORAGE_SIZE \
570 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
572 struct bin_tree_storage_t
574 struct bin_tree_storage_t *next;
575 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
577 typedef struct bin_tree_storage_t bin_tree_storage_t;
579 #define CONTEXT_WORD 1
580 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
581 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
582 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
584 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
585 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
586 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
587 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
588 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
590 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
591 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
592 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
593 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
595 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
596 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
597 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
598 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
599 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
601 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
602 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
603 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
604 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
605 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
611 re_node_set non_eps_nodes;
612 re_node_set inveclosure;
613 re_node_set *entrance_nodes;
614 struct re_dfastate_t **trtable, **word_trtable;
615 unsigned int context : 4;
616 unsigned int halt : 1;
617 /* If this state can accept `multi byte'.
618 Note that we refer to multibyte characters, and multi character
619 collating elements as `multi byte'. */
620 unsigned int accept_mb : 1;
621 /* If this state has backreference node(s). */
622 unsigned int has_backref : 1;
623 unsigned int has_constraint : 1;
625 typedef struct re_dfastate_t re_dfastate_t;
627 struct re_state_table_entry
631 re_dfastate_t **array;
634 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
640 re_dfastate_t **array;
643 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
648 Idx str_idx; /* The position NODE match at. */
650 } re_sub_match_last_t;
652 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
653 And information about the node, whose type is OP_CLOSE_SUBEXP,
654 corresponding to NODE is stored in LASTS. */
661 Idx alasts; /* Allocation size of LASTS. */
662 Idx nlasts; /* The number of LASTS. */
663 re_sub_match_last_t **lasts;
664 } re_sub_match_top_t;
666 struct re_backref_cache_entry
674 unsigned short int eps_reachable_subexps_map;
679 /* The string object corresponding to the input string. */
681 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
686 /* EFLAGS of the argument of regexec. */
688 /* Where the matching ends. */
691 /* The state log used by the matcher. */
692 re_dfastate_t **state_log;
694 /* Back reference cache. */
697 struct re_backref_cache_entry *bkref_ents;
701 re_sub_match_top_t **sub_tops;
702 } re_match_context_t;
706 re_dfastate_t **sifted_states;
707 re_dfastate_t **limited_states;
713 struct re_fail_stack_ent_t
718 re_node_set eps_via_nodes;
721 struct re_fail_stack_t
725 struct re_fail_stack_ent_t *stack;
736 re_node_set *eclosures;
737 re_node_set *inveclosures;
738 struct re_state_table_entry *state_table;
739 re_dfastate_t *init_state;
740 re_dfastate_t *init_state_word;
741 re_dfastate_t *init_state_nl;
742 re_dfastate_t *init_state_begbuf;
743 bin_tree_t *str_tree;
744 bin_tree_storage_t *str_tree_storage;
745 re_bitset_ptr_t sb_char;
746 int str_tree_storage_idx;
748 /* number of subexpressions `re_nsub' is in regex_t. */
749 re_hashval_t state_hash_mask;
751 Idx nbackref; /* The number of backreference in this dfa. */
753 /* Bitmap expressing which backreference is used. */
754 bitset_word used_bkref_map;
755 bitset_word completed_bkref_map;
757 unsigned int has_plural_match : 1;
758 /* If this dfa has "multibyte node", which is a backreference or
759 a node which can accept multibyte character or multi character
760 collating element. */
761 unsigned int has_mb_node : 1;
762 unsigned int is_utf8 : 1;
763 unsigned int map_notascii : 1;
764 unsigned int word_ops_used : 1;
772 __libc_lock_define (, lock)
775 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
776 #define re_node_set_remove(set,id) \
777 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
778 #define re_node_set_empty(p) ((p)->nelem = 0)
779 #define re_node_set_free(set) re_free ((set)->elems)
781 static void free_state (re_dfastate_t *state) internal_function;
795 bracket_elem_type type;
805 /* Inline functions for bitset operation. */
808 bitset_set (bitset set, Idx i)
810 set[i / BITSET_WORD_BITS] |= (bitset_word) 1 << i % BITSET_WORD_BITS;
814 bitset_clear (bitset set, Idx i)
816 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word) 1 << i % BITSET_WORD_BITS);
820 bitset_contain (const bitset set, Idx i)
822 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
826 bitset_empty (bitset set)
828 memset (set, 0, sizeof (bitset));
832 bitset_set_all (bitset set)
834 memset (set, -1, sizeof (bitset_word) * (SBC_MAX / BITSET_WORD_BITS));
835 if (SBC_MAX % BITSET_WORD_BITS != 0)
836 set[BITSET_WORDS - 1] =
837 ((bitset_word) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
841 bitset_copy (bitset dest, const bitset src)
843 memcpy (dest, src, sizeof (bitset));
847 bitset_not (bitset set)
850 for (i = 0; i < SBC_MAX / BITSET_WORD_BITS; ++i)
852 if (SBC_MAX % BITSET_WORD_BITS != 0)
853 set[BITSET_WORDS - 1] =
854 ((((bitset_word) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
855 & ~set[BITSET_WORDS - 1]);
859 bitset_merge (bitset dest, const bitset src)
862 for (i = 0; i < BITSET_WORDS; ++i)
867 bitset_mask (bitset dest, const bitset src)
870 for (i = 0; i < BITSET_WORDS; ++i)
874 #if defined RE_ENABLE_I18N
875 /* Inline functions for re_string. */
877 internal_function __attribute ((pure))
878 re_string_char_size_at (const re_string_t *pstr, Idx idx)
881 if (pstr->mb_cur_max == 1)
883 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
884 if (pstr->wcs[idx + byte_idx] != WEOF)
890 internal_function __attribute ((pure))
891 re_string_wchar_at (const re_string_t *pstr, Idx idx)
893 if (pstr->mb_cur_max == 1)
894 return (wint_t) pstr->mbs[idx];
895 return (wint_t) pstr->wcs[idx];
899 internal_function __attribute ((pure))
900 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
903 const unsigned char *p, *extra;
904 const int32_t *table, *indirect;
906 # include <locale/weight.h>
907 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
911 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
912 extra = (const unsigned char *)
913 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
914 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
915 _NL_COLLATE_INDIRECTMB);
918 return p - pstr->mbs - idx;
924 #endif /* RE_ENABLE_I18N */
926 #endif /* _REGEX_INTERNAL_H */