1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002, 2003, 2004, 2005, 2006 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 # define BE(expr, val) __builtin_expect (expr, val)
94 # define BE(expr, val) (expr)
97 /* Number of single byte character. */
100 #define COLL_ELEM_LEN_MAX 8
102 /* The character which represents newline. */
103 #define NEWLINE_CHAR '\n'
104 #define WIDE_NEWLINE_CHAR L'\n'
106 /* Rename to standard API for using out of glibc. */
108 # define __wctype wctype
109 # define __iswctype iswctype
110 # define __btowc btowc
112 # define __mempcpy mempcpy
114 # define __wcrtomb wcrtomb
115 # define __regfree regfree
116 # define attribute_hidden
117 #endif /* not _LIBC */
119 #if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
120 # define __attribute(arg) __attribute__ (arg)
122 # define __attribute(arg)
125 extern const char __re_error_msgid[] attribute_hidden;
126 extern const size_t __re_error_msgid_idx[] attribute_hidden;
128 typedef __re_idx_t Idx;
130 /* Special return value for failure to match. */
131 #define REG_MISSING ((Idx) -1)
133 /* Special return value for internal error. */
134 #define REG_ERROR ((Idx) -2)
136 /* Test whether N is a valid index, and is not one of the above. */
137 #ifdef _REGEX_LARGE_OFFSETS
138 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
140 # define REG_VALID_INDEX(n) (0 <= (n))
143 /* Test whether N is a valid nonzero index. */
144 #ifdef _REGEX_LARGE_OFFSETS
145 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
147 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
150 /* A hash value, suitable for computing hash tables. */
151 typedef __re_size_t re_hashval_t;
153 /* An integer used to represent a set of bits. It must be unsigned,
154 and must be at least as wide as unsigned int. */
155 typedef unsigned long int bitset_word;
157 /* Maximum value of a bitset word. It must be useful in preprocessor
158 contexts, and must be consistent with bitset_word. */
159 #define BITSET_WORD_MAX ULONG_MAX
161 /* Number of bits in a bitset word. Avoid greater-than-32-bit
162 integers and unconditional shifts by more than 31 bits, as they're
164 #if BITSET_WORD_MAX == 0xffffffff
165 # define BITSET_WORD_BITS 32
166 #elif BITSET_WORD_MAX >> 31 >> 5 == 1
167 # define BITSET_WORD_BITS 36
168 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
169 # define BITSET_WORD_BITS 48
170 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
171 # define BITSET_WORD_BITS 60
172 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
173 # define BITSET_WORD_BITS 64
174 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
175 # define BITSET_WORD_BITS 72
176 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
177 # define BITSET_WORD_BITS 128
178 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
179 # define BITSET_WORD_BITS 256
180 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
181 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
182 # if BITSET_WORD_BITS <= SBC_MAX
183 # error "Invalid SBC_MAX"
185 #elif BITSET_WORD_MAX == (0xffffffff + 2) * 0xffffffff
186 /* Work around a bug in 64-bit PGC (before version 6.1-2), where the
187 preprocessor mishandles large unsigned values as if they were signed. */
188 # define BITSET_WORD_BITS 64
190 # error "Add case for new bitset_word size"
193 /* Number of bitset words in a bitset. */
194 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
196 typedef bitset_word bitset[BITSET_WORDS];
197 typedef bitset_word *re_bitset_ptr_t;
198 typedef const bitset_word *re_const_bitset_ptr_t;
200 #define PREV_WORD_CONSTRAINT 0x0001
201 #define PREV_NOTWORD_CONSTRAINT 0x0002
202 #define NEXT_WORD_CONSTRAINT 0x0004
203 #define NEXT_NOTWORD_CONSTRAINT 0x0008
204 #define PREV_NEWLINE_CONSTRAINT 0x0010
205 #define NEXT_NEWLINE_CONSTRAINT 0x0020
206 #define PREV_BEGBUF_CONSTRAINT 0x0040
207 #define NEXT_ENDBUF_CONSTRAINT 0x0080
208 #define WORD_DELIM_CONSTRAINT 0x0100
209 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
213 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
214 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
215 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
216 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
217 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
218 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
219 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
220 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
221 WORD_DELIM = WORD_DELIM_CONSTRAINT,
222 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
236 /* Node type, These are used by token, node, tree. */
242 #ifdef RE_ENABLE_I18N
245 #endif /* RE_ENABLE_I18N */
247 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
248 when the debugger shows values of this enum type. */
249 #define EPSILON_BIT 8
250 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
251 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
252 OP_ALT = EPSILON_BIT | 2,
253 OP_DUP_ASTERISK = EPSILON_BIT | 3,
254 ANCHOR = EPSILON_BIT | 4,
256 /* Tree type, these are used only by tree. */
260 /* Token type, these are used only by token. */
272 OP_CLOSE_EQUIV_CLASS,
283 #ifdef RE_ENABLE_I18N
286 /* Multibyte characters. */
289 /* Collating symbols. */
294 /* Equivalence classes. */
296 int32_t *equiv_classes;
299 /* Range expressions. */
301 uint32_t *range_starts;
302 uint32_t *range_ends;
303 # else /* not _LIBC */
304 wchar_t *range_starts;
306 # endif /* not _LIBC */
308 /* Character classes. */
309 wctype_t *char_classes;
311 /* If this character set is the non-matching list. */
312 unsigned int non_match : 1;
314 /* # of multibyte characters. */
317 /* # of collating symbols. */
320 /* # of equivalence classes. */
323 /* # of range expressions. */
326 /* # of character classes. */
329 #endif /* RE_ENABLE_I18N */
335 unsigned char c; /* for CHARACTER */
336 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
337 #ifdef RE_ENABLE_I18N
338 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
339 #endif /* RE_ENABLE_I18N */
340 Idx idx; /* for BACK_REF */
341 re_context_type ctx_type; /* for ANCHOR */
344 re_token_type_t type : 8;
346 re_token_type_t type;
348 unsigned int constraint : 10; /* context constraint */
349 unsigned int duplicated : 1;
350 unsigned int opt_subexp : 1;
351 #ifdef RE_ENABLE_I18N
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;
357 unsigned int word_char : 1;
360 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
364 /* Indicate the raw buffer which is the original string passed as an
365 argument of regexec(), re_search(), etc.. */
366 const unsigned char *raw_mbs;
367 /* Store the multibyte string. In case of "case insensitive mode" like
368 REG_ICASE, upper cases of the string are stored, otherwise MBS points
369 the same address that RAW_MBS points. */
371 #ifdef RE_ENABLE_I18N
372 /* Store the wide character string which is corresponding to MBS. */
377 /* Index in RAW_MBS. Each character mbs[i] corresponds to
378 raw_mbs[raw_mbs_idx + i]. */
380 /* The length of the valid characters in the buffers. */
382 /* The corresponding number of bytes in raw_mbs array. */
384 /* The length of the buffers MBS and WCS. */
386 /* The index in MBS, which is updated by re_string_fetch_byte. */
388 /* length of RAW_MBS array. */
390 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
392 /* End of the buffer may be shorter than its length in the cases such
393 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
396 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
399 /* The context of mbs[0]. We store the context independently, since
400 the context of mbs[0] may be different from raw_mbs[0], which is
401 the beginning of the input string. */
402 unsigned int tip_context;
403 /* The translation passed as a part of an argument of re_compile_pattern. */
404 unsigned REG_TRANSLATE_TYPE trans;
405 /* Copy of re_dfa_t's word_char. */
406 re_const_bitset_ptr_t word_char;
407 /* true if REG_ICASE. */
409 unsigned char is_utf8;
410 unsigned char map_notascii;
411 unsigned char mbs_allocated;
412 unsigned char offsets_needed;
413 unsigned char newline_anchor;
414 unsigned char word_ops_used;
417 typedef struct re_string_t re_string_t;
421 typedef struct re_dfa_t re_dfa_t;
425 # define internal_function __attribute ((regparm (3), stdcall))
427 # define internal_function
431 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
434 #ifdef RE_ENABLE_I18N
435 static void build_wcs_buffer (re_string_t *pstr) internal_function;
436 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
438 #endif /* RE_ENABLE_I18N */
439 static void build_upper_buffer (re_string_t *pstr) internal_function;
440 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
441 static unsigned int re_string_context_at (const re_string_t *input,
443 internal_function __attribute ((pure));
445 #define re_string_peek_byte(pstr, offset) \
446 ((pstr)->mbs[(pstr)->cur_idx + offset])
447 #define re_string_fetch_byte(pstr) \
448 ((pstr)->mbs[(pstr)->cur_idx++])
449 #define re_string_first_byte(pstr, idx) \
450 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
451 #define re_string_is_single_byte_char(pstr, idx) \
452 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
453 || (pstr)->wcs[(idx) + 1] != WEOF))
454 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
455 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
456 #define re_string_get_buffer(pstr) ((pstr)->mbs)
457 #define re_string_length(pstr) ((pstr)->len)
458 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
459 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
460 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
466 /* The OS usually guarantees only one guard page at the bottom of the stack,
467 and a page size can be as small as 4096 bytes. So we cannot safely
468 allocate anything larger than 4096 bytes. Also care for the possibility
469 of a few compiler-allocated temporary stack slots. */
470 # define __libc_use_alloca(n) ((n) < 4032)
472 /* alloca is implemented with malloc, so just use malloc. */
473 # define __libc_use_alloca(n) 0
477 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
478 #define re_xmalloc(t,n) ((t *) re_xnmalloc (n, sizeof (t)))
479 #define re_calloc(t,n) ((t *) calloc (n, sizeof (t)))
480 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
481 #define re_xrealloc(p,t,n) ((t *) re_xnrealloc (p, n, sizeof (t)))
482 #define re_x2realloc(p,t,pn) ((t *) re_x2nrealloc (p, pn, sizeof (t)))
483 #define re_free(p) free (p)
486 # define SIZE_MAX ((size_t) -1)
489 /* Return true if an array of N objects, each of size S, cannot exist
490 due to size arithmetic overflow. S must be nonzero. */
492 re_alloc_oversized (size_t n, size_t s)
494 return BE (SIZE_MAX / s < n, 0);
497 /* Return true if an array of (2 * N + 1) objects, each of size S,
498 cannot exist due to size arithmetic overflow. S must be nonzero. */
500 re_x2alloc_oversized (size_t n, size_t s)
502 return BE ((SIZE_MAX / s - 1) / 2 < n, 0);
505 /* Allocate an array of N objects, each with S bytes of memory,
506 dynamically, with error checking. S must be nonzero. */
508 re_xnmalloc (size_t n, size_t s)
510 return re_alloc_oversized (n, s) ? NULL : malloc (n * s);
513 /* Change the size of an allocated block of memory P to an array of N
514 objects each of S bytes, with error checking. S must be nonzero. */
516 re_xnrealloc (void *p, size_t n, size_t s)
518 return re_alloc_oversized (n, s) ? NULL : realloc (p, n * s);
521 /* Reallocate a block of memory P to an array of (2 * (*PN) + 1)
522 objects each of S bytes, with error checking. S must be nonzero.
523 If the allocation is successful, set *PN to the new allocation
524 count and return the resulting pointer. Otherwise, return
527 re_x2nrealloc (void *p, size_t *pn, size_t s)
529 if (re_x2alloc_oversized (*pn, s))
533 /* Add 1 in case *PN is zero. */
534 size_t n1 = 2 * *pn + 1;
535 p = realloc (p, n1 * s);
536 if (BE (p != NULL, 1))
544 struct bin_tree_t *parent;
545 struct bin_tree_t *left;
546 struct bin_tree_t *right;
547 struct bin_tree_t *first;
548 struct bin_tree_t *next;
552 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
553 Otherwise `type' indicate the type of this node. */
556 typedef struct bin_tree_t bin_tree_t;
558 #define BIN_TREE_STORAGE_SIZE \
559 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
561 struct bin_tree_storage_t
563 struct bin_tree_storage_t *next;
564 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
566 typedef struct bin_tree_storage_t bin_tree_storage_t;
568 #define CONTEXT_WORD 1
569 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
570 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
571 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
573 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
574 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
575 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
576 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
577 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
579 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
580 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
581 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
582 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
584 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
585 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
586 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
587 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
588 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
590 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
591 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
592 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
593 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
594 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
600 re_node_set non_eps_nodes;
601 re_node_set inveclosure;
602 re_node_set *entrance_nodes;
603 struct re_dfastate_t **trtable, **word_trtable;
604 unsigned int context : 4;
605 unsigned int halt : 1;
606 /* If this state can accept `multi byte'.
607 Note that we refer to multibyte characters, and multi character
608 collating elements as `multi byte'. */
609 unsigned int accept_mb : 1;
610 /* If this state has backreference node(s). */
611 unsigned int has_backref : 1;
612 unsigned int has_constraint : 1;
614 typedef struct re_dfastate_t re_dfastate_t;
616 struct re_state_table_entry
620 re_dfastate_t **array;
623 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
629 re_dfastate_t **array;
632 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
637 Idx str_idx; /* The position NODE match at. */
639 } re_sub_match_last_t;
641 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
642 And information about the node, whose type is OP_CLOSE_SUBEXP,
643 corresponding to NODE is stored in LASTS. */
650 Idx alasts; /* Allocation size of LASTS. */
651 Idx nlasts; /* The number of LASTS. */
652 re_sub_match_last_t **lasts;
653 } re_sub_match_top_t;
655 struct re_backref_cache_entry
663 unsigned short int eps_reachable_subexps_map;
668 /* The string object corresponding to the input string. */
670 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
675 /* EFLAGS of the argument of regexec. */
677 /* Where the matching ends. */
680 /* The state log used by the matcher. */
681 re_dfastate_t **state_log;
683 /* Back reference cache. */
686 struct re_backref_cache_entry *bkref_ents;
690 re_sub_match_top_t **sub_tops;
691 } re_match_context_t;
695 re_dfastate_t **sifted_states;
696 re_dfastate_t **limited_states;
702 struct re_fail_stack_ent_t
707 re_node_set eps_via_nodes;
710 struct re_fail_stack_t
714 struct re_fail_stack_ent_t *stack;
725 re_node_set *eclosures;
726 re_node_set *inveclosures;
727 struct re_state_table_entry *state_table;
728 re_dfastate_t *init_state;
729 re_dfastate_t *init_state_word;
730 re_dfastate_t *init_state_nl;
731 re_dfastate_t *init_state_begbuf;
732 bin_tree_t *str_tree;
733 bin_tree_storage_t *str_tree_storage;
734 re_bitset_ptr_t sb_char;
735 int str_tree_storage_idx;
737 /* number of subexpressions `re_nsub' is in regex_t. */
738 re_hashval_t state_hash_mask;
740 Idx nbackref; /* The number of backreference in this dfa. */
742 /* Bitmap expressing which backreference is used. */
743 bitset_word used_bkref_map;
744 bitset_word completed_bkref_map;
746 unsigned int has_plural_match : 1;
747 /* If this dfa has "multibyte node", which is a backreference or
748 a node which can accept multibyte character or multi character
749 collating element. */
750 unsigned int has_mb_node : 1;
751 unsigned int is_utf8 : 1;
752 unsigned int map_notascii : 1;
753 unsigned int word_ops_used : 1;
761 __libc_lock_define (, lock)
764 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
765 #define re_node_set_remove(set,id) \
766 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
767 #define re_node_set_empty(p) ((p)->nelem = 0)
768 #define re_node_set_free(set) re_free ((set)->elems)
770 static void free_state (re_dfastate_t *state) internal_function;
784 bracket_elem_type type;
794 /* Inline functions for bitset operation. */
797 bitset_set (bitset set, Idx i)
799 set[i / BITSET_WORD_BITS] |= (bitset_word) 1 << i % BITSET_WORD_BITS;
803 bitset_clear (bitset set, Idx i)
805 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word) 1 << i % BITSET_WORD_BITS);
809 bitset_contain (const bitset set, Idx i)
811 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
815 bitset_empty (bitset set)
817 memset (set, 0, sizeof (bitset));
821 bitset_set_all (bitset set)
823 memset (set, -1, sizeof (bitset_word) * (SBC_MAX / BITSET_WORD_BITS));
824 if (SBC_MAX % BITSET_WORD_BITS != 0)
825 set[BITSET_WORDS - 1] =
826 ((bitset_word) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
830 bitset_copy (bitset dest, const bitset src)
832 memcpy (dest, src, sizeof (bitset));
836 bitset_not (bitset set)
839 for (i = 0; i < SBC_MAX / BITSET_WORD_BITS; ++i)
841 if (SBC_MAX % BITSET_WORD_BITS != 0)
842 set[BITSET_WORDS - 1] =
843 ((((bitset_word) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
844 & ~set[BITSET_WORDS - 1]);
848 bitset_merge (bitset dest, const bitset src)
851 for (i = 0; i < BITSET_WORDS; ++i)
856 bitset_mask (bitset dest, const bitset src)
859 for (i = 0; i < BITSET_WORDS; ++i)
863 #if defined RE_ENABLE_I18N
864 /* Inline functions for re_string. */
866 internal_function __attribute ((pure))
867 re_string_char_size_at (const re_string_t *pstr, Idx idx)
870 if (pstr->mb_cur_max == 1)
872 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
873 if (pstr->wcs[idx + byte_idx] != WEOF)
879 internal_function __attribute ((pure))
880 re_string_wchar_at (const re_string_t *pstr, Idx idx)
882 if (pstr->mb_cur_max == 1)
883 return (wint_t) pstr->mbs[idx];
884 return (wint_t) pstr->wcs[idx];
888 internal_function __attribute ((pure))
889 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
892 const unsigned char *p, *extra;
893 const int32_t *table, *indirect;
895 # include <locale/weight.h>
896 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
900 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
901 extra = (const unsigned char *)
902 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
903 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
904 _NL_COLLATE_INDIRECTMB);
907 return p - pstr->mbs - idx;
913 #endif /* RE_ENABLE_I18N */
915 #endif /* _REGEX_INTERNAL_H */