1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010 Free
3 Software Foundation, Inc.
4 This file is part of the GNU C Library.
5 Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
7 This program is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3, or (at your option)
12 This program is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License along
18 with this program; if not, write to the Free Software Foundation,
19 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
21 #ifndef _REGEX_INTERNAL_H
22 #define _REGEX_INTERNAL_H 1
33 # include "localcharset.h"
35 #if defined HAVE_LOCALE_H || defined _LIBC
43 # include <bits/libc-lock.h>
45 # define __libc_lock_init(NAME) do { } while (0)
46 # define __libc_lock_lock(NAME) do { } while (0)
47 # define __libc_lock_unlock(NAME) do { } while (0)
50 /* In case that the system doesn't have isblank(). */
51 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
52 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
56 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
57 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
58 # include <locale/localeinfo.h>
59 # include <locale/elem-hash.h>
60 # include <locale/coll-lookup.h>
64 /* This is for other GNU distributions with internationalized messages. */
65 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
69 # define gettext(msgid) \
70 INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES)
73 # define gettext(msgid) (msgid)
77 /* This define is so xgettext can find the internationalizable
79 # define gettext_noop(String) String
82 /* For loser systems without the definition. */
84 # define SIZE_MAX ((size_t) -1)
87 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_ISWCTYPE && HAVE_WCSCOLL) || _LIBC
88 # define RE_ENABLE_I18N
92 # define BE(expr, val) __builtin_expect (expr, val)
94 # define BE(expr, val) (expr)
100 /* Number of ASCII characters. */
101 #define ASCII_CHARS 0x80
103 /* Number of single byte characters. */
104 #define SBC_MAX (UCHAR_MAX + 1)
106 #define COLL_ELEM_LEN_MAX 8
108 /* The character which represents newline. */
109 #define NEWLINE_CHAR '\n'
110 #define WIDE_NEWLINE_CHAR L'\n'
112 /* Rename to standard API for using out of glibc. */
114 # define __wctype wctype
115 # define __iswctype iswctype
116 # define __btowc btowc
117 # define __wcrtomb wcrtomb
118 # define __mbrtowc mbrtowc
119 # define __regfree regfree
120 # define attribute_hidden
121 #endif /* not _LIBC */
123 #if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
124 # define __attribute(arg) __attribute__ (arg)
126 # define __attribute(arg)
129 typedef __re_idx_t Idx
;
131 /* Special return value for failure to match. */
132 #define REG_MISSING ((Idx) -1)
134 /* Special return value for internal error. */
135 #define REG_ERROR ((Idx) -2)
137 /* Test whether N is a valid index, and is not one of the above. */
138 #ifdef _REGEX_LARGE_OFFSETS
139 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
141 # define REG_VALID_INDEX(n) (0 <= (n))
144 /* Test whether N is a valid nonzero index. */
145 #ifdef _REGEX_LARGE_OFFSETS
146 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
148 # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
151 /* A hash value, suitable for computing hash tables. */
152 typedef __re_size_t re_hashval_t
;
154 /* An integer used to represent a set of bits. It must be unsigned,
155 and must be at least as wide as unsigned int. */
156 typedef unsigned long int bitset_word_t
;
157 /* All bits set in a bitset_word_t. */
158 #define BITSET_WORD_MAX ULONG_MAX
160 /* Number of bits in a bitset_word_t. For portability to hosts with
161 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
162 instead, deduce it directly from BITSET_WORD_MAX. Avoid
163 greater-than-32-bit integers and unconditional shifts by more than
164 31 bits, as they're not portable. */
165 #if BITSET_WORD_MAX == 0xffffffffUL
166 # define BITSET_WORD_BITS 32
167 #elif BITSET_WORD_MAX >> 31 >> 4 == 1
168 # define BITSET_WORD_BITS 36
169 #elif BITSET_WORD_MAX >> 31 >> 16 == 1
170 # define BITSET_WORD_BITS 48
171 #elif BITSET_WORD_MAX >> 31 >> 28 == 1
172 # define BITSET_WORD_BITS 60
173 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
174 # define BITSET_WORD_BITS 64
175 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
176 # define BITSET_WORD_BITS 72
177 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
178 # define BITSET_WORD_BITS 128
179 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
180 # define BITSET_WORD_BITS 256
181 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
182 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
183 # if BITSET_WORD_BITS <= SBC_MAX
184 # error "Invalid SBC_MAX"
187 # error "Add case for new bitset_word_t size"
190 /* Number of bitset_word_t values in a bitset_t. */
191 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
193 typedef bitset_word_t bitset_t
[BITSET_WORDS
];
194 typedef bitset_word_t
*re_bitset_ptr_t
;
195 typedef const bitset_word_t
*re_const_bitset_ptr_t
;
197 #define PREV_WORD_CONSTRAINT 0x0001
198 #define PREV_NOTWORD_CONSTRAINT 0x0002
199 #define NEXT_WORD_CONSTRAINT 0x0004
200 #define NEXT_NOTWORD_CONSTRAINT 0x0008
201 #define PREV_NEWLINE_CONSTRAINT 0x0010
202 #define NEXT_NEWLINE_CONSTRAINT 0x0020
203 #define PREV_BEGBUF_CONSTRAINT 0x0040
204 #define NEXT_ENDBUF_CONSTRAINT 0x0080
205 #define WORD_DELIM_CONSTRAINT 0x0100
206 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
210 INSIDE_WORD
= PREV_WORD_CONSTRAINT
| NEXT_WORD_CONSTRAINT
,
211 WORD_FIRST
= PREV_NOTWORD_CONSTRAINT
| NEXT_WORD_CONSTRAINT
,
212 WORD_LAST
= PREV_WORD_CONSTRAINT
| NEXT_NOTWORD_CONSTRAINT
,
213 INSIDE_NOTWORD
= PREV_NOTWORD_CONSTRAINT
| NEXT_NOTWORD_CONSTRAINT
,
214 LINE_FIRST
= PREV_NEWLINE_CONSTRAINT
,
215 LINE_LAST
= NEXT_NEWLINE_CONSTRAINT
,
216 BUF_FIRST
= PREV_BEGBUF_CONSTRAINT
,
217 BUF_LAST
= NEXT_ENDBUF_CONSTRAINT
,
218 WORD_DELIM
= WORD_DELIM_CONSTRAINT
,
219 NOT_WORD_DELIM
= NOT_WORD_DELIM_CONSTRAINT
233 /* Node type, These are used by token, node, tree. */
239 #ifdef RE_ENABLE_I18N
242 #endif /* RE_ENABLE_I18N */
244 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
245 when the debugger shows values of this enum type. */
246 #define EPSILON_BIT 8
247 OP_OPEN_SUBEXP
= EPSILON_BIT
| 0,
248 OP_CLOSE_SUBEXP
= EPSILON_BIT
| 1,
249 OP_ALT
= EPSILON_BIT
| 2,
250 OP_DUP_ASTERISK
= EPSILON_BIT
| 3,
251 ANCHOR
= EPSILON_BIT
| 4,
253 /* Tree type, these are used only by tree. */
257 /* Token type, these are used only by token. */
269 OP_CLOSE_EQUIV_CLASS
,
280 #ifdef RE_ENABLE_I18N
283 /* Multibyte characters. */
286 /* Collating symbols. */
291 /* Equivalence classes. */
293 int32_t *equiv_classes
;
296 /* Range expressions. */
298 uint32_t *range_starts
;
299 uint32_t *range_ends
;
300 # else /* not _LIBC */
301 wchar_t *range_starts
;
303 # endif /* not _LIBC */
305 /* Character classes. */
306 wctype_t *char_classes
;
308 /* If this character set is the non-matching list. */
309 unsigned int non_match
: 1;
311 /* # of multibyte characters. */
314 /* # of collating symbols. */
317 /* # of equivalence classes. */
320 /* # of range expressions. */
323 /* # of character classes. */
326 #endif /* RE_ENABLE_I18N */
332 unsigned char c
; /* for CHARACTER */
333 re_bitset_ptr_t sbcset
; /* for SIMPLE_BRACKET */
334 #ifdef RE_ENABLE_I18N
335 re_charset_t
*mbcset
; /* for COMPLEX_BRACKET */
336 #endif /* RE_ENABLE_I18N */
337 Idx idx
; /* for BACK_REF */
338 re_context_type ctx_type
; /* for ANCHOR */
340 #if __GNUC__ >= 2 && !__STRICT_ANSI__
341 re_token_type_t type
: 8;
343 re_token_type_t type
;
345 unsigned int constraint
: 10; /* context constraint */
346 unsigned int duplicated
: 1;
347 unsigned int opt_subexp
: 1;
348 #ifdef RE_ENABLE_I18N
349 unsigned int accept_mb
: 1;
350 /* These 2 bits can be moved into the union if needed (e.g. if running out
351 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
352 unsigned int mb_partial
: 1;
354 unsigned int word_char
: 1;
357 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
361 /* Indicate the raw buffer which is the original string passed as an
362 argument of regexec(), re_search(), etc.. */
363 const unsigned char *raw_mbs
;
364 /* Store the multibyte string. In case of "case insensitive mode" like
365 REG_ICASE, upper cases of the string are stored, otherwise MBS points
366 the same address that RAW_MBS points. */
368 #ifdef RE_ENABLE_I18N
369 /* Store the wide character string which is corresponding to MBS. */
374 /* Index in RAW_MBS. Each character mbs[i] corresponds to
375 raw_mbs[raw_mbs_idx + i]. */
377 /* The length of the valid characters in the buffers. */
379 /* The corresponding number of bytes in raw_mbs array. */
381 /* The length of the buffers MBS and WCS. */
383 /* The index in MBS, which is updated by re_string_fetch_byte. */
385 /* length of RAW_MBS array. */
387 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
389 /* End of the buffer may be shorter than its length in the cases such
390 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
393 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
396 /* The context of mbs[0]. We store the context independently, since
397 the context of mbs[0] may be different from raw_mbs[0], which is
398 the beginning of the input string. */
399 unsigned int tip_context
;
400 /* The translation passed as a part of an argument of re_compile_pattern. */
401 RE_TRANSLATE_TYPE trans
;
402 /* Copy of re_dfa_t's word_char. */
403 re_const_bitset_ptr_t word_char
;
404 /* true if REG_ICASE. */
406 unsigned char is_utf8
;
407 unsigned char map_notascii
;
408 unsigned char mbs_allocated
;
409 unsigned char offsets_needed
;
410 unsigned char newline_anchor
;
411 unsigned char word_ops_used
;
414 typedef struct re_string_t re_string_t
;
417 # if defined __i386__ && !defined __EMX__
418 # define internal_function __attribute ((regparm (3), stdcall))
420 # define internal_function
424 static reg_errcode_t
re_string_realloc_buffers (re_string_t
*pstr
,
427 #ifdef RE_ENABLE_I18N
428 static void build_wcs_buffer (re_string_t
*pstr
) internal_function
;
429 static reg_errcode_t
build_wcs_upper_buffer (re_string_t
*pstr
)
431 #endif /* RE_ENABLE_I18N */
432 static void build_upper_buffer (re_string_t
*pstr
) internal_function
;
433 static void re_string_translate_buffer (re_string_t
*pstr
) internal_function
;
434 static unsigned int re_string_context_at (const re_string_t
*input
, Idx idx
,
436 internal_function
__attribute ((pure
));
437 #define re_string_peek_byte(pstr, offset) \
438 ((pstr)->mbs[(pstr)->cur_idx + offset])
439 #define re_string_fetch_byte(pstr) \
440 ((pstr)->mbs[(pstr)->cur_idx++])
441 #define re_string_first_byte(pstr, idx) \
442 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
443 #define re_string_is_single_byte_char(pstr, idx) \
444 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
445 || (pstr)->wcs[(idx) + 1] != WEOF))
446 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
447 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
448 #define re_string_get_buffer(pstr) ((pstr)->mbs)
449 #define re_string_length(pstr) ((pstr)->len)
450 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
451 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
452 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
458 /* The OS usually guarantees only one guard page at the bottom of the stack,
459 and a page size can be as small as 4096 bytes. So we cannot safely
460 allocate anything larger than 4096 bytes. Also care for the possibility
461 of a few compiler-allocated temporary stack slots. */
462 # define __libc_use_alloca(n) ((n) < 4032)
464 /* alloca is implemented with malloc, so just use malloc. */
465 # define __libc_use_alloca(n) 0
467 # define alloca(n) malloc (n)
472 # define MAX(a,b) ((a) < (b) ? (b) : (a))
475 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
476 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
477 #define re_free(p) free (p)
481 struct bin_tree_t
*parent
;
482 struct bin_tree_t
*left
;
483 struct bin_tree_t
*right
;
484 struct bin_tree_t
*first
;
485 struct bin_tree_t
*next
;
489 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
490 Otherwise `type' indicate the type of this node. */
493 typedef struct bin_tree_t bin_tree_t
;
495 #define BIN_TREE_STORAGE_SIZE \
496 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
498 struct bin_tree_storage_t
500 struct bin_tree_storage_t
*next
;
501 bin_tree_t data
[BIN_TREE_STORAGE_SIZE
];
503 typedef struct bin_tree_storage_t bin_tree_storage_t
;
505 #define CONTEXT_WORD 1
506 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
507 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
508 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
510 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
511 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
512 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
513 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
514 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
516 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
517 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
518 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
519 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
521 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
522 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
523 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
524 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
525 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
527 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
528 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
529 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
530 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
531 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
537 re_node_set non_eps_nodes
;
538 re_node_set inveclosure
;
539 re_node_set
*entrance_nodes
;
540 struct re_dfastate_t
**trtable
, **word_trtable
;
541 unsigned int context
: 4;
542 unsigned int halt
: 1;
543 /* If this state can accept `multi byte'.
544 Note that we refer to multibyte characters, and multi character
545 collating elements as `multi byte'. */
546 unsigned int accept_mb
: 1;
547 /* If this state has backreference node(s). */
548 unsigned int has_backref
: 1;
549 unsigned int has_constraint
: 1;
551 typedef struct re_dfastate_t re_dfastate_t
;
553 struct re_state_table_entry
557 re_dfastate_t
**array
;
560 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
566 re_dfastate_t
**array
;
569 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
574 Idx str_idx
; /* The position NODE match at. */
576 } re_sub_match_last_t
;
578 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
579 And information about the node, whose type is OP_CLOSE_SUBEXP,
580 corresponding to NODE is stored in LASTS. */
587 Idx alasts
; /* Allocation size of LASTS. */
588 Idx nlasts
; /* The number of LASTS. */
589 re_sub_match_last_t
**lasts
;
590 } re_sub_match_top_t
;
592 struct re_backref_cache_entry
600 unsigned short int eps_reachable_subexps_map
;
605 /* The string object corresponding to the input string. */
607 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
608 const re_dfa_t
*const dfa
;
612 /* EFLAGS of the argument of regexec. */
614 /* Where the matching ends. */
617 /* The state log used by the matcher. */
618 re_dfastate_t
**state_log
;
620 /* Back reference cache. */
623 struct re_backref_cache_entry
*bkref_ents
;
627 re_sub_match_top_t
**sub_tops
;
628 } re_match_context_t
;
632 re_dfastate_t
**sifted_states
;
633 re_dfastate_t
**limited_states
;
639 struct re_fail_stack_ent_t
644 re_node_set eps_via_nodes
;
647 struct re_fail_stack_t
651 struct re_fail_stack_ent_t
*stack
;
662 re_node_set
*eclosures
;
663 re_node_set
*inveclosures
;
664 struct re_state_table_entry
*state_table
;
665 re_dfastate_t
*init_state
;
666 re_dfastate_t
*init_state_word
;
667 re_dfastate_t
*init_state_nl
;
668 re_dfastate_t
*init_state_begbuf
;
669 bin_tree_t
*str_tree
;
670 bin_tree_storage_t
*str_tree_storage
;
671 re_bitset_ptr_t sb_char
;
672 int str_tree_storage_idx
;
674 /* number of subexpressions `re_nsub' is in regex_t. */
675 re_hashval_t state_hash_mask
;
677 Idx nbackref
; /* The number of backreference in this dfa. */
679 /* Bitmap expressing which backreference is used. */
680 bitset_word_t used_bkref_map
;
681 bitset_word_t completed_bkref_map
;
683 unsigned int has_plural_match
: 1;
684 /* If this dfa has "multibyte node", which is a backreference or
685 a node which can accept multibyte character or multi character
686 collating element. */
687 unsigned int has_mb_node
: 1;
688 unsigned int is_utf8
: 1;
689 unsigned int map_notascii
: 1;
690 unsigned int word_ops_used
: 1;
699 __libc_lock_define (, lock
)
703 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
704 #define re_node_set_remove(set,id) \
705 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
706 #define re_node_set_empty(p) ((p)->nelem = 0)
707 #define re_node_set_free(set) re_free ((set)->elems)
721 bracket_elem_type type
;
731 /* Inline functions for bitset_t operation. */
734 bitset_set (bitset_t set
, Idx i
)
736 set
[i
/ BITSET_WORD_BITS
] |= (bitset_word_t
) 1 << i
% BITSET_WORD_BITS
;
740 bitset_clear (bitset_t set
, Idx i
)
742 set
[i
/ BITSET_WORD_BITS
] &= ~ ((bitset_word_t
) 1 << i
% BITSET_WORD_BITS
);
746 bitset_contain (const bitset_t set
, Idx i
)
748 return (set
[i
/ BITSET_WORD_BITS
] >> i
% BITSET_WORD_BITS
) & 1;
752 bitset_empty (bitset_t set
)
754 memset (set
, '\0', sizeof (bitset_t
));
758 bitset_set_all (bitset_t set
)
760 memset (set
, -1, sizeof (bitset_word_t
) * (SBC_MAX
/ BITSET_WORD_BITS
));
761 if (SBC_MAX
% BITSET_WORD_BITS
!= 0)
762 set
[BITSET_WORDS
- 1] =
763 ((bitset_word_t
) 1 << SBC_MAX
% BITSET_WORD_BITS
) - 1;
767 bitset_copy (bitset_t dest
, const bitset_t src
)
769 memcpy (dest
, src
, sizeof (bitset_t
));
773 bitset_not (bitset_t set
)
776 for (bitset_i
= 0; bitset_i
< SBC_MAX
/ BITSET_WORD_BITS
; ++bitset_i
)
777 set
[bitset_i
] = ~set
[bitset_i
];
778 if (SBC_MAX
% BITSET_WORD_BITS
!= 0)
779 set
[BITSET_WORDS
- 1] =
780 ((((bitset_word_t
) 1 << SBC_MAX
% BITSET_WORD_BITS
) - 1)
781 & ~set
[BITSET_WORDS
- 1]);
785 bitset_merge (bitset_t dest
, const bitset_t src
)
788 for (bitset_i
= 0; bitset_i
< BITSET_WORDS
; ++bitset_i
)
789 dest
[bitset_i
] |= src
[bitset_i
];
793 bitset_mask (bitset_t dest
, const bitset_t src
)
796 for (bitset_i
= 0; bitset_i
< BITSET_WORDS
; ++bitset_i
)
797 dest
[bitset_i
] &= src
[bitset_i
];
800 #ifdef RE_ENABLE_I18N
801 /* Inline functions for re_string. */
803 internal_function
__attribute ((pure
))
804 re_string_char_size_at (const re_string_t
*pstr
, Idx idx
)
807 if (pstr
->mb_cur_max
== 1)
809 for (byte_idx
= 1; idx
+ byte_idx
< pstr
->valid_len
; ++byte_idx
)
810 if (pstr
->wcs
[idx
+ byte_idx
] != WEOF
)
816 internal_function
__attribute ((pure
))
817 re_string_wchar_at (const re_string_t
*pstr
, Idx idx
)
819 if (pstr
->mb_cur_max
== 1)
820 return (wint_t) pstr
->mbs
[idx
];
821 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
;
832 # include <locale/weight.h>
833 uint_fast32_t nrules
= _NL_CURRENT_WORD (LC_COLLATE
, _NL_COLLATE_NRULES
);
837 table
= (const int32_t *) _NL_CURRENT (LC_COLLATE
, _NL_COLLATE_TABLEMB
);
838 extra
= (const unsigned char *)
839 _NL_CURRENT (LC_COLLATE
, _NL_COLLATE_EXTRAMB
);
840 indirect
= (const int32_t *) _NL_CURRENT (LC_COLLATE
,
841 _NL_COLLATE_INDIRECTMB
);
844 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 */