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
30 #if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET || defined _LIBC
31 # include <langinfo.h>
33 #if defined HAVE_LOCALE_H || defined _LIBC
36 #if defined HAVE_WCHAR_H || defined _LIBC
38 #endif /* HAVE_WCHAR_H || _LIBC */
39 #if defined HAVE_WCTYPE_H || defined _LIBC
41 #endif /* HAVE_WCTYPE_H || _LIBC */
43 # include <bits/libc-lock.h>
45 # define __libc_lock_define(CLASS,NAME)
46 # define __libc_lock_init(NAME) do { } while (0)
47 # define __libc_lock_lock(NAME) do { } while (0)
48 # define __libc_lock_unlock(NAME) do { } while (0)
51 /* In case that the system doesn't have isblank(). */
52 #if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank
53 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
57 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
58 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
59 # include <locale/localeinfo.h>
60 # include <locale/elem-hash.h>
61 # include <locale/coll-lookup.h>
65 /* This is for other GNU distributions with internationalized messages. */
66 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
70 # define gettext(msgid) \
71 INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES)
74 # define gettext(msgid) (msgid)
78 /* This define is so xgettext can find the internationalizable
80 # define gettext_noop(String) String
83 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC
84 # define RE_ENABLE_I18N
88 # if defined __GNUC__ && defined __GNUC_MINOR__
89 # define __GNUC_PREREQ(maj, min) \
90 ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min))
92 # define __GNUC_PREREQ(maj, min) 0
96 #if !__GNUC_PREREQ (3, 1)
97 # define always_inline
101 # define BE(expr, val) __builtin_expect (expr, val)
103 # define BE(expr, val) (expr)
108 /* Number of single byte character. */
111 #define COLL_ELEM_LEN_MAX 8
113 /* The character which represents newline. */
114 #define NEWLINE_CHAR '\n'
115 #define WIDE_NEWLINE_CHAR L'\n'
117 /* Rename to standard API for using out of glibc. */
119 # define __wctype wctype
120 # define __iswctype iswctype
121 # define __btowc btowc
123 # define __mempcpy mempcpy
125 # define __wcrtomb wcrtomb
126 # define __regfree regfree
127 # define attribute_hidden
128 #endif /* not _LIBC */
131 # define __attribute(arg) __attribute__ (arg)
133 # define __attribute(arg)
136 extern const char __re_error_msgid[] attribute_hidden;
137 extern const size_t __re_error_msgid_idx[] attribute_hidden;
139 typedef __re_idx_t Idx;
141 /* Special return value for failure to match. */
142 #define REG_MISSING ((Idx) -1)
144 /* Special return value for internal error. */
145 #define REG_ERROR ((Idx) -2)
147 /* Test whether N is a valid index, and is not one of the above. */
148 #ifdef _REGEX_LARGE_OFFSETS
149 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
151 # define REG_VALID_INDEX(n) (0 <= (n))
154 /* Test whether N is a valid nonzero index. */
155 #ifdef _REGEX_LARGE_OFFSETS
156 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
158 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
161 /* A hash value, suitable for computing hash tables. */
162 typedef __re_size_t re_hashval_t;
164 /* An integer used to represent a set of bits. It must be unsigned,
165 and must be at least as wide as unsigned int. */
166 typedef unsigned long int bitset_word;
168 /* Maximum value of a bitset word. It must be useful in preprocessor
169 contexts, and must be consistent with bitset_word. */
170 #define BITSET_WORD_MAX ULONG_MAX
172 /* Number of bits in a bitset word. Avoid greater-than-32-bit
173 integers and unconditional shifts by more than 31 bits, as they're
175 #if BITSET_WORD_MAX == 0xffffffff
176 # define BITSET_WORD_BITS 32
177 #elif BITSET_WORD_MAX >> 31 >> 5 == 1
178 # define BITSET_WORD_BITS 36
179 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
180 # define BITSET_WORD_BITS 48
181 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
182 # define BITSET_WORD_BITS 60
183 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
184 # define BITSET_WORD_BITS 64
185 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
186 # define BITSET_WORD_BITS 72
187 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
188 # define BITSET_WORD_BITS 128
189 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
190 # define BITSET_WORD_BITS 256
191 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
192 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
193 # if BITSET_WORD_BITS <= SBC_MAX
194 # error "Invalid SBC_MAX"
197 # error "Add case for new bitset_word size"
200 /* Number of bitset words in a bitset. */
201 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
203 typedef bitset_word bitset[BITSET_WORDS];
204 typedef bitset_word *re_bitset_ptr_t;
205 typedef const bitset_word *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. */
249 #ifdef RE_ENABLE_I18N
252 #endif /* RE_ENABLE_I18N */
254 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
255 when the debugger shows values of this enum type. */
256 #define EPSILON_BIT 8
257 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
258 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
259 OP_ALT = EPSILON_BIT | 2,
260 OP_DUP_ASTERISK = EPSILON_BIT | 3,
261 ANCHOR = EPSILON_BIT | 4,
263 /* Tree type, these are used only by tree. */
267 /* Token type, these are used only by token. */
279 OP_CLOSE_EQUIV_CLASS,
290 #ifdef RE_ENABLE_I18N
293 /* Multibyte characters. */
296 /* Collating symbols. */
301 /* Equivalence classes. */
303 int32_t *equiv_classes;
306 /* Range expressions. */
308 uint32_t *range_starts;
309 uint32_t *range_ends;
310 # else /* not _LIBC */
311 wchar_t *range_starts;
313 # endif /* not _LIBC */
315 /* Character classes. */
316 wctype_t *char_classes;
318 /* If this character set is the non-matching list. */
319 unsigned int non_match : 1;
321 /* # of multibyte characters. */
324 /* # of collating symbols. */
327 /* # of equivalence classes. */
330 /* # of range expressions. */
333 /* # of character classes. */
336 #endif /* RE_ENABLE_I18N */
342 unsigned char c; /* for CHARACTER */
343 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
344 #ifdef RE_ENABLE_I18N
345 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
346 #endif /* RE_ENABLE_I18N */
347 Idx idx; /* for BACK_REF */
348 re_context_type ctx_type; /* for ANCHOR */
351 re_token_type_t type : 8;
353 re_token_type_t type;
355 unsigned int constraint : 10; /* context constraint */
356 unsigned int duplicated : 1;
357 unsigned int opt_subexp : 1;
358 #ifdef RE_ENABLE_I18N
359 unsigned int accept_mb : 1;
360 /* These 2 bits can be moved into the union if needed (e.g. if running out
361 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
362 unsigned int mb_partial : 1;
364 unsigned int word_char : 1;
367 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
371 /* Indicate the raw buffer which is the original string passed as an
372 argument of regexec(), re_search(), etc.. */
373 const unsigned char *raw_mbs;
374 /* Store the multibyte string. In case of "case insensitive mode" like
375 REG_ICASE, upper cases of the string are stored, otherwise MBS points
376 the same address that RAW_MBS points. */
378 #ifdef RE_ENABLE_I18N
379 /* Store the wide character string which is corresponding to MBS. */
384 /* Index in RAW_MBS. Each character mbs[i] corresponds to
385 raw_mbs[raw_mbs_idx + i]. */
387 /* The length of the valid characters in the buffers. */
389 /* The corresponding number of bytes in raw_mbs array. */
391 /* The length of the buffers MBS and WCS. */
393 /* The index in MBS, which is updated by re_string_fetch_byte. */
395 /* length of RAW_MBS array. */
397 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
399 /* End of the buffer may be shorter than its length in the cases such
400 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
403 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
406 /* The context of mbs[0]. We store the context independently, since
407 the context of mbs[0] may be different from raw_mbs[0], which is
408 the beginning of the input string. */
409 unsigned int tip_context;
410 /* The translation passed as a part of an argument of re_compile_pattern. */
411 unsigned REG_TRANSLATE_TYPE trans;
412 /* Copy of re_dfa_t's word_char. */
413 re_const_bitset_ptr_t word_char;
414 /* true if REG_ICASE. */
416 unsigned char is_utf8;
417 unsigned char map_notascii;
418 unsigned char mbs_allocated;
419 unsigned char offsets_needed;
420 unsigned char newline_anchor;
421 unsigned char word_ops_used;
424 typedef struct re_string_t re_string_t;
428 typedef struct re_dfa_t re_dfa_t;
432 # define internal_function __attribute ((regparm (3), stdcall))
434 # define internal_function
438 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
441 #ifdef RE_ENABLE_I18N
442 static void build_wcs_buffer (re_string_t *pstr) internal_function;
443 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
445 #endif /* RE_ENABLE_I18N */
446 static void build_upper_buffer (re_string_t *pstr) internal_function;
447 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
448 static unsigned int re_string_context_at (const re_string_t *input,
450 internal_function __attribute ((pure));
452 #define re_string_peek_byte(pstr, offset) \
453 ((pstr)->mbs[(pstr)->cur_idx + offset])
454 #define re_string_fetch_byte(pstr) \
455 ((pstr)->mbs[(pstr)->cur_idx++])
456 #define re_string_first_byte(pstr, idx) \
457 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
458 #define re_string_is_single_byte_char(pstr, idx) \
459 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
460 || (pstr)->wcs[(idx) + 1] != WEOF))
461 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
462 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
463 #define re_string_get_buffer(pstr) ((pstr)->mbs)
464 #define re_string_length(pstr) ((pstr)->len)
465 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
466 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
467 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
473 /* The OS usually guarantees only one guard page at the bottom of the stack,
474 and a page size can be as small as 4096 bytes. So we cannot safely
475 allocate anything larger than 4096 bytes. Also care for the possibility
476 of a few compiler-allocated temporary stack slots. */
477 # define __libc_use_alloca(n) ((n) < 4032)
479 /* alloca is implemented with malloc, so just use malloc. */
480 # define __libc_use_alloca(n) 0
484 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
485 #define re_xmalloc(t,n) ((t *) re_xnmalloc (n, sizeof (t)))
486 #define re_calloc(t,n) ((t *) calloc (n, sizeof (t)))
487 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
488 #define re_xrealloc(p,t,n) ((t *) re_xnrealloc (p, n, sizeof (t)))
489 #define re_x2realloc(p,t,pn) ((t *) re_x2nrealloc (p, pn, sizeof (t)))
490 #define re_free(p) free (p)
493 # define SIZE_MAX ((size_t) -1)
496 /* Return true if an array of N objects, each of size S, cannot exist
497 due to size arithmetic overflow. S must be nonzero. */
499 re_alloc_oversized (size_t n, size_t s)
501 return BE (SIZE_MAX / s < n, 0);
504 /* Return true if an array of (2 * N + 1) objects, each of size S,
505 cannot exist due to size arithmetic overflow. S must be nonzero. */
507 re_x2alloc_oversized (size_t n, size_t s)
509 return BE ((SIZE_MAX / s - 1) / 2 < n, 0);
512 /* Allocate an array of N objects, each with S bytes of memory,
513 dynamically, with error checking. S must be nonzero. */
515 re_xnmalloc (size_t n, size_t s)
517 return re_alloc_oversized (n, s) ? NULL : malloc (n * s);
520 /* Change the size of an allocated block of memory P to an array of N
521 objects each of S bytes, with error checking. S must be nonzero. */
523 re_xnrealloc (void *p, size_t n, size_t s)
525 return re_alloc_oversized (n, s) ? NULL : realloc (p, n * s);
528 /* Reallocate a block of memory P to an array of (2 * (*PN) + 1)
529 objects each of S bytes, with error checking. S must be nonzero.
530 If the allocation is successful, set *PN to the new allocation
531 count and return the resulting pointer. Otherwise, return
534 re_x2nrealloc (void *p, size_t *pn, size_t s)
536 if (re_x2alloc_oversized (*pn, s))
540 /* Add 1 in case *PN is zero. */
541 size_t n1 = 2 * *pn + 1;
542 p = realloc (p, n1 * s);
543 if (BE (p != NULL, 1))
551 struct bin_tree_t *parent;
552 struct bin_tree_t *left;
553 struct bin_tree_t *right;
554 struct bin_tree_t *first;
555 struct bin_tree_t *next;
559 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
560 Otherwise `type' indicate the type of this node. */
563 typedef struct bin_tree_t bin_tree_t;
565 #define BIN_TREE_STORAGE_SIZE \
566 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
568 struct bin_tree_storage_t
570 struct bin_tree_storage_t *next;
571 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
573 typedef struct bin_tree_storage_t bin_tree_storage_t;
575 #define CONTEXT_WORD 1
576 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
577 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
578 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
580 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
581 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
582 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
583 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
584 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
586 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
587 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
588 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
589 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
591 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
592 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
593 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
594 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
595 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
597 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
598 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
599 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
600 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
601 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
607 re_node_set non_eps_nodes;
608 re_node_set inveclosure;
609 re_node_set *entrance_nodes;
610 struct re_dfastate_t **trtable, **word_trtable;
611 unsigned int context : 4;
612 unsigned int halt : 1;
613 /* If this state can accept `multi byte'.
614 Note that we refer to multibyte characters, and multi character
615 collating elements as `multi byte'. */
616 unsigned int accept_mb : 1;
617 /* If this state has backreference node(s). */
618 unsigned int has_backref : 1;
619 unsigned int has_constraint : 1;
621 typedef struct re_dfastate_t re_dfastate_t;
623 struct re_state_table_entry
627 re_dfastate_t **array;
630 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
636 re_dfastate_t **array;
639 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
644 Idx str_idx; /* The position NODE match at. */
646 } re_sub_match_last_t;
648 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
649 And information about the node, whose type is OP_CLOSE_SUBEXP,
650 corresponding to NODE is stored in LASTS. */
657 Idx alasts; /* Allocation size of LASTS. */
658 Idx nlasts; /* The number of LASTS. */
659 re_sub_match_last_t **lasts;
660 } re_sub_match_top_t;
662 struct re_backref_cache_entry
670 unsigned short int eps_reachable_subexps_map;
675 /* The string object corresponding to the input string. */
677 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
682 /* EFLAGS of the argument of regexec. */
684 /* Where the matching ends. */
687 /* The state log used by the matcher. */
688 re_dfastate_t **state_log;
690 /* Back reference cache. */
693 struct re_backref_cache_entry *bkref_ents;
697 re_sub_match_top_t **sub_tops;
698 } re_match_context_t;
702 re_dfastate_t **sifted_states;
703 re_dfastate_t **limited_states;
709 struct re_fail_stack_ent_t
714 re_node_set eps_via_nodes;
717 struct re_fail_stack_t
721 struct re_fail_stack_ent_t *stack;
732 re_node_set *eclosures;
733 re_node_set *inveclosures;
734 struct re_state_table_entry *state_table;
735 re_dfastate_t *init_state;
736 re_dfastate_t *init_state_word;
737 re_dfastate_t *init_state_nl;
738 re_dfastate_t *init_state_begbuf;
739 bin_tree_t *str_tree;
740 bin_tree_storage_t *str_tree_storage;
741 re_bitset_ptr_t sb_char;
742 int str_tree_storage_idx;
744 /* number of subexpressions `re_nsub' is in regex_t. */
745 re_hashval_t state_hash_mask;
747 Idx nbackref; /* The number of backreference in this dfa. */
749 /* Bitmap expressing which backreference is used. */
750 bitset_word used_bkref_map;
751 bitset_word completed_bkref_map;
753 unsigned int has_plural_match : 1;
754 /* If this dfa has "multibyte node", which is a backreference or
755 a node which can accept multibyte character or multi character
756 collating element. */
757 unsigned int has_mb_node : 1;
758 unsigned int is_utf8 : 1;
759 unsigned int map_notascii : 1;
760 unsigned int word_ops_used : 1;
768 __libc_lock_define (, lock)
771 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
772 #define re_node_set_remove(set,id) \
773 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
774 #define re_node_set_empty(p) ((p)->nelem = 0)
775 #define re_node_set_free(set) re_free ((set)->elems)
777 static void free_state (re_dfastate_t *state) internal_function;
791 bracket_elem_type type;
801 /* Inline functions for bitset operation. */
804 bitset_set (bitset set, Idx i)
806 set[i / BITSET_WORD_BITS] |= (bitset_word) 1 << i % BITSET_WORD_BITS;
810 bitset_clear (bitset set, Idx i)
812 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word) 1 << i % BITSET_WORD_BITS);
816 bitset_contain (const bitset set, Idx i)
818 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
822 bitset_empty (bitset set)
824 memset (set, 0, sizeof (bitset));
828 bitset_set_all (bitset set)
830 memset (set, -1, sizeof (bitset_word) * (SBC_MAX / BITSET_WORD_BITS));
831 if (SBC_MAX % BITSET_WORD_BITS != 0)
832 set[BITSET_WORDS - 1] =
833 ((bitset_word) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
837 bitset_copy (bitset dest, const bitset src)
839 memcpy (dest, src, sizeof (bitset));
843 bitset_not (bitset set)
846 for (i = 0; i < SBC_MAX / BITSET_WORD_BITS; ++i)
848 if (SBC_MAX % BITSET_WORD_BITS != 0)
849 set[BITSET_WORDS - 1] =
850 ((((bitset_word) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
851 & ~set[BITSET_WORDS - 1]);
855 bitset_merge (bitset dest, const bitset src)
858 for (i = 0; i < BITSET_WORDS; ++i)
863 bitset_mask (bitset dest, const bitset src)
866 for (i = 0; i < BITSET_WORDS; ++i)
870 #if defined RE_ENABLE_I18N
871 /* Inline functions for re_string. */
873 internal_function __attribute ((pure))
874 re_string_char_size_at (const re_string_t *pstr, Idx idx)
877 if (pstr->mb_cur_max == 1)
879 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
880 if (pstr->wcs[idx + byte_idx] != WEOF)
886 internal_function __attribute ((pure))
887 re_string_wchar_at (const re_string_t *pstr, Idx idx)
889 if (pstr->mb_cur_max == 1)
890 return (wint_t) pstr->mbs[idx];
891 return (wint_t) pstr->wcs[idx];
895 internal_function __attribute ((pure))
896 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
899 const unsigned char *p, *extra;
900 const int32_t *table, *indirect;
902 # include <locale/weight.h>
903 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
907 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
908 extra = (const unsigned char *)
909 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
910 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
911 _NL_COLLATE_INDIRECTMB);
914 return p - pstr->mbs - idx;
920 #endif /* RE_ENABLE_I18N */
922 #endif /* _REGEX_INTERNAL_H */