Clean up __exit_thread.
[glibc.git] / posix / regex_internal.h
blob75c390f8c8b624993a14c502272dde2357d91ccc
1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002-2014 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4 Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
6 The GNU C Library is free software; you can redistribute it and/or
7 modify it under the terms of the GNU Lesser General Public
8 License as published by the Free Software Foundation; either
9 version 2.1 of the License, or (at your option) any later version.
11 The GNU C Library 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 GNU
14 Lesser General Public License for more details.
16 You should have received a copy of the GNU Lesser General Public
17 License along with the GNU C Library; if not, see
18 <http://www.gnu.org/licenses/>. */
20 #ifndef _REGEX_INTERNAL_H
21 #define _REGEX_INTERNAL_H 1
23 #include <assert.h>
24 #include <ctype.h>
25 #include <stdio.h>
26 #include <stdlib.h>
27 #include <string.h>
29 #if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET || defined _LIBC
30 # include <langinfo.h>
31 #endif
32 #if defined HAVE_LOCALE_H || defined _LIBC
33 # include <locale.h>
34 #endif
35 #if defined HAVE_WCHAR_H || defined _LIBC
36 # include <wchar.h>
37 #endif /* HAVE_WCHAR_H || _LIBC */
38 #if defined HAVE_WCTYPE_H || defined _LIBC
39 # include <wctype.h>
40 #endif /* HAVE_WCTYPE_H || _LIBC */
41 #if defined HAVE_STDBOOL_H || defined _LIBC
42 # include <stdbool.h>
43 #endif /* HAVE_STDBOOL_H || _LIBC */
44 #if defined HAVE_STDINT_H || defined _LIBC
45 # include <stdint.h>
46 #endif /* HAVE_STDINT_H || _LIBC */
47 #if defined _LIBC
48 # include <bits/libc-lock.h>
49 #else
50 # define __libc_lock_define(CLASS,NAME)
51 # define __libc_lock_init(NAME) do { } while (0)
52 # define __libc_lock_lock(NAME) do { } while (0)
53 # define __libc_lock_unlock(NAME) do { } while (0)
54 #endif
56 /* In case that the system doesn't have isblank(). */
57 #if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank
58 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
59 #endif
61 #ifdef _LIBC
62 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
63 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
64 # include <locale/localeinfo.h>
65 # include <locale/elem-hash.h>
66 # include <locale/coll-lookup.h>
67 # endif
68 #endif
70 /* This is for other GNU distributions with internationalized messages. */
71 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
72 # include <libintl.h>
73 # ifdef _LIBC
74 # undef gettext
75 # define gettext(msgid) \
76 __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
77 # endif
78 #else
79 # define gettext(msgid) (msgid)
80 #endif
82 #ifndef gettext_noop
83 /* This define is so xgettext can find the internationalizable
84 strings. */
85 # define gettext_noop(String) String
86 #endif
88 /* For loser systems without the definition. */
89 #ifndef SIZE_MAX
90 # define SIZE_MAX ((size_t) -1)
91 #endif
93 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC
94 # define RE_ENABLE_I18N
95 #endif
97 #if __GNUC__ >= 3
98 # define BE(expr, val) __builtin_expect (expr, val)
99 #else
100 # define BE(expr, val) (expr)
101 #endif
103 /* Number of single byte character. */
104 #define SBC_MAX 256
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. */
113 #ifndef _LIBC
114 # define __wctype wctype
115 # define __iswctype iswctype
116 # define __btowc btowc
117 # define __mbrtowc mbrtowc
118 # define __mempcpy mempcpy
119 # define __wcrtomb wcrtomb
120 # define __regfree regfree
121 # define attribute_hidden
122 #endif /* not _LIBC */
124 #if __GNUC__ < 3 + (__GNUC_MINOR__ < 1)
125 # define __attribute__(arg)
126 #endif
128 extern const char __re_error_msgid[] attribute_hidden;
129 extern const size_t __re_error_msgid_idx[] attribute_hidden;
131 /* An integer used to represent a set of bits. It must be unsigned,
132 and must be at least as wide as unsigned int. */
133 typedef unsigned long int bitset_word_t;
134 /* All bits set in a bitset_word_t. */
135 #define BITSET_WORD_MAX ULONG_MAX
136 /* Number of bits in a bitset_word_t. */
137 #define BITSET_WORD_BITS (sizeof (bitset_word_t) * CHAR_BIT)
138 /* Number of bitset_word_t in a bit_set. */
139 #define BITSET_WORDS (SBC_MAX / BITSET_WORD_BITS)
140 typedef bitset_word_t bitset_t[BITSET_WORDS];
141 typedef bitset_word_t *re_bitset_ptr_t;
142 typedef const bitset_word_t *re_const_bitset_ptr_t;
144 #define bitset_set(set,i) \
145 (set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS)
146 #define bitset_clear(set,i) \
147 (set[i / BITSET_WORD_BITS] &= ~((bitset_word_t) 1 << i % BITSET_WORD_BITS))
148 #define bitset_contain(set,i) \
149 (set[i / BITSET_WORD_BITS] & ((bitset_word_t) 1 << i % BITSET_WORD_BITS))
150 #define bitset_empty(set) memset (set, '\0', sizeof (bitset_t))
151 #define bitset_set_all(set) memset (set, '\xff', sizeof (bitset_t))
152 #define bitset_copy(dest,src) memcpy (dest, src, sizeof (bitset_t))
154 #define PREV_WORD_CONSTRAINT 0x0001
155 #define PREV_NOTWORD_CONSTRAINT 0x0002
156 #define NEXT_WORD_CONSTRAINT 0x0004
157 #define NEXT_NOTWORD_CONSTRAINT 0x0008
158 #define PREV_NEWLINE_CONSTRAINT 0x0010
159 #define NEXT_NEWLINE_CONSTRAINT 0x0020
160 #define PREV_BEGBUF_CONSTRAINT 0x0040
161 #define NEXT_ENDBUF_CONSTRAINT 0x0080
162 #define WORD_DELIM_CONSTRAINT 0x0100
163 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
165 typedef enum
167 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
168 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
169 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
170 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
171 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
172 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
173 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
174 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
175 WORD_DELIM = WORD_DELIM_CONSTRAINT,
176 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
177 } re_context_type;
179 typedef struct
181 int alloc;
182 int nelem;
183 int *elems;
184 } re_node_set;
186 typedef enum
188 NON_TYPE = 0,
190 /* Node type, These are used by token, node, tree. */
191 CHARACTER = 1,
192 END_OF_RE = 2,
193 SIMPLE_BRACKET = 3,
194 OP_BACK_REF = 4,
195 OP_PERIOD = 5,
196 #ifdef RE_ENABLE_I18N
197 COMPLEX_BRACKET = 6,
198 OP_UTF8_PERIOD = 7,
199 #endif /* RE_ENABLE_I18N */
201 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
202 when the debugger shows values of this enum type. */
203 #define EPSILON_BIT 8
204 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
205 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
206 OP_ALT = EPSILON_BIT | 2,
207 OP_DUP_ASTERISK = EPSILON_BIT | 3,
208 ANCHOR = EPSILON_BIT | 4,
210 /* Tree type, these are used only by tree. */
211 CONCAT = 16,
212 SUBEXP = 17,
214 /* Token type, these are used only by token. */
215 OP_DUP_PLUS = 18,
216 OP_DUP_QUESTION,
217 OP_OPEN_BRACKET,
218 OP_CLOSE_BRACKET,
219 OP_CHARSET_RANGE,
220 OP_OPEN_DUP_NUM,
221 OP_CLOSE_DUP_NUM,
222 OP_NON_MATCH_LIST,
223 OP_OPEN_COLL_ELEM,
224 OP_CLOSE_COLL_ELEM,
225 OP_OPEN_EQUIV_CLASS,
226 OP_CLOSE_EQUIV_CLASS,
227 OP_OPEN_CHAR_CLASS,
228 OP_CLOSE_CHAR_CLASS,
229 OP_WORD,
230 OP_NOTWORD,
231 OP_SPACE,
232 OP_NOTSPACE,
233 BACK_SLASH
235 } re_token_type_t;
237 #ifdef RE_ENABLE_I18N
238 typedef struct
240 /* Multibyte characters. */
241 wchar_t *mbchars;
243 /* Collating symbols. */
244 # ifdef _LIBC
245 int32_t *coll_syms;
246 # endif
248 /* Equivalence classes. */
249 # ifdef _LIBC
250 int32_t *equiv_classes;
251 # endif
253 /* Range expressions. */
254 # ifdef _LIBC
255 uint32_t *range_starts;
256 uint32_t *range_ends;
257 # else /* not _LIBC */
258 wchar_t *range_starts;
259 wchar_t *range_ends;
260 # endif /* not _LIBC */
262 /* Character classes. */
263 wctype_t *char_classes;
265 /* If this character set is the non-matching list. */
266 unsigned int non_match : 1;
268 /* # of multibyte characters. */
269 int nmbchars;
271 /* # of collating symbols. */
272 int ncoll_syms;
274 /* # of equivalence classes. */
275 int nequiv_classes;
277 /* # of range expressions. */
278 int nranges;
280 /* # of character classes. */
281 int nchar_classes;
282 } re_charset_t;
283 #endif /* RE_ENABLE_I18N */
285 typedef struct
287 union
289 unsigned char c; /* for CHARACTER */
290 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
291 #ifdef RE_ENABLE_I18N
292 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
293 #endif /* RE_ENABLE_I18N */
294 int idx; /* for BACK_REF */
295 re_context_type ctx_type; /* for ANCHOR */
296 } opr;
297 #if __GNUC__ >= 2
298 re_token_type_t type : 8;
299 #else
300 re_token_type_t type;
301 #endif
302 unsigned int constraint : 10; /* context constraint */
303 unsigned int duplicated : 1;
304 unsigned int opt_subexp : 1;
305 #ifdef RE_ENABLE_I18N
306 unsigned int accept_mb : 1;
307 /* These 2 bits can be moved into the union if needed (e.g. if running out
308 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
309 unsigned int mb_partial : 1;
310 #endif
311 unsigned int word_char : 1;
312 } re_token_t;
314 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
316 struct re_string_t
318 /* Indicate the raw buffer which is the original string passed as an
319 argument of regexec(), re_search(), etc.. */
320 const unsigned char *raw_mbs;
321 /* Store the multibyte string. In case of "case insensitive mode" like
322 REG_ICASE, upper cases of the string are stored, otherwise MBS points
323 the same address that RAW_MBS points. */
324 unsigned char *mbs;
325 #ifdef RE_ENABLE_I18N
326 /* Store the wide character string which is corresponding to MBS. */
327 wint_t *wcs;
328 int *offsets;
329 mbstate_t cur_state;
330 #endif
331 /* Index in RAW_MBS. Each character mbs[i] corresponds to
332 raw_mbs[raw_mbs_idx + i]. */
333 int raw_mbs_idx;
334 /* The length of the valid characters in the buffers. */
335 int valid_len;
336 /* The corresponding number of bytes in raw_mbs array. */
337 int valid_raw_len;
338 /* The length of the buffers MBS and WCS. */
339 int bufs_len;
340 /* The index in MBS, which is updated by re_string_fetch_byte. */
341 int cur_idx;
342 /* length of RAW_MBS array. */
343 int raw_len;
344 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
345 int len;
346 /* End of the buffer may be shorter than its length in the cases such
347 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
348 instead of LEN. */
349 int raw_stop;
350 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
351 int stop;
353 /* The context of mbs[0]. We store the context independently, since
354 the context of mbs[0] may be different from raw_mbs[0], which is
355 the beginning of the input string. */
356 unsigned int tip_context;
357 /* The translation passed as a part of an argument of re_compile_pattern. */
358 RE_TRANSLATE_TYPE trans;
359 /* Copy of re_dfa_t's word_char. */
360 re_const_bitset_ptr_t word_char;
361 /* 1 if REG_ICASE. */
362 unsigned char icase;
363 unsigned char is_utf8;
364 unsigned char map_notascii;
365 unsigned char mbs_allocated;
366 unsigned char offsets_needed;
367 unsigned char newline_anchor;
368 unsigned char word_ops_used;
369 int mb_cur_max;
371 typedef struct re_string_t re_string_t;
374 struct re_dfa_t;
375 typedef struct re_dfa_t re_dfa_t;
377 #ifndef _LIBC
378 # ifdef __i386__
379 # define internal_function __attribute__ ((regparm (3), stdcall))
380 # else
381 # define internal_function
382 # endif
383 #endif
385 #ifndef NOT_IN_libc
386 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
387 int new_buf_len)
388 internal_function;
389 # ifdef RE_ENABLE_I18N
390 static void build_wcs_buffer (re_string_t *pstr) internal_function;
391 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
392 internal_function;
393 # endif /* RE_ENABLE_I18N */
394 static void build_upper_buffer (re_string_t *pstr) internal_function;
395 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
396 static unsigned int re_string_context_at (const re_string_t *input, int idx,
397 int eflags)
398 internal_function __attribute__ ((pure));
399 #endif
400 #define re_string_peek_byte(pstr, offset) \
401 ((pstr)->mbs[(pstr)->cur_idx + offset])
402 #define re_string_fetch_byte(pstr) \
403 ((pstr)->mbs[(pstr)->cur_idx++])
404 #define re_string_first_byte(pstr, idx) \
405 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
406 #define re_string_is_single_byte_char(pstr, idx) \
407 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
408 || (pstr)->wcs[(idx) + 1] != WEOF))
409 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
410 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
411 #define re_string_get_buffer(pstr) ((pstr)->mbs)
412 #define re_string_length(pstr) ((pstr)->len)
413 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
414 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
415 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
417 #include <alloca.h>
419 #ifndef _LIBC
420 # if HAVE_ALLOCA
421 /* The OS usually guarantees only one guard page at the bottom of the stack,
422 and a page size can be as small as 4096 bytes. So we cannot safely
423 allocate anything larger than 4096 bytes. Also care for the possibility
424 of a few compiler-allocated temporary stack slots. */
425 # define __libc_use_alloca(n) ((n) < 4032)
426 # else
427 /* alloca is implemented with malloc, so just use malloc. */
428 # define __libc_use_alloca(n) 0
429 # endif
430 #endif
432 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
433 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
434 #define re_free(p) free (p)
436 struct bin_tree_t
438 struct bin_tree_t *parent;
439 struct bin_tree_t *left;
440 struct bin_tree_t *right;
441 struct bin_tree_t *first;
442 struct bin_tree_t *next;
444 re_token_t token;
446 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
447 Otherwise `type' indicate the type of this node. */
448 int node_idx;
450 typedef struct bin_tree_t bin_tree_t;
452 #define BIN_TREE_STORAGE_SIZE \
453 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
455 struct bin_tree_storage_t
457 struct bin_tree_storage_t *next;
458 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
460 typedef struct bin_tree_storage_t bin_tree_storage_t;
462 #define CONTEXT_WORD 1
463 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
464 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
465 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
467 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
468 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
469 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
470 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
471 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
473 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
474 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
475 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
476 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
478 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
479 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
480 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
481 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
482 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
484 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
485 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
486 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
487 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
488 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
490 struct re_dfastate_t
492 unsigned int hash;
493 re_node_set nodes;
494 re_node_set non_eps_nodes;
495 re_node_set inveclosure;
496 re_node_set *entrance_nodes;
497 struct re_dfastate_t **trtable, **word_trtable;
498 unsigned int context : 4;
499 unsigned int halt : 1;
500 /* If this state can accept `multi byte'.
501 Note that we refer to multibyte characters, and multi character
502 collating elements as `multi byte'. */
503 unsigned int accept_mb : 1;
504 /* If this state has backreference node(s). */
505 unsigned int has_backref : 1;
506 unsigned int has_constraint : 1;
508 typedef struct re_dfastate_t re_dfastate_t;
510 struct re_state_table_entry
512 int num;
513 int alloc;
514 re_dfastate_t **array;
517 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
519 typedef struct
521 int next_idx;
522 int alloc;
523 re_dfastate_t **array;
524 } state_array_t;
526 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
528 typedef struct
530 int node;
531 int str_idx; /* The position NODE match at. */
532 state_array_t path;
533 } re_sub_match_last_t;
535 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
536 And information about the node, whose type is OP_CLOSE_SUBEXP,
537 corresponding to NODE is stored in LASTS. */
539 typedef struct
541 int str_idx;
542 int node;
543 state_array_t *path;
544 int alasts; /* Allocation size of LASTS. */
545 int nlasts; /* The number of LASTS. */
546 re_sub_match_last_t **lasts;
547 } re_sub_match_top_t;
549 struct re_backref_cache_entry
551 int node;
552 int str_idx;
553 int subexp_from;
554 int subexp_to;
555 char more;
556 char unused;
557 unsigned short int eps_reachable_subexps_map;
560 typedef struct
562 /* The string object corresponding to the input string. */
563 re_string_t input;
564 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
565 const re_dfa_t *const dfa;
566 #else
567 const re_dfa_t *dfa;
568 #endif
569 /* EFLAGS of the argument of regexec. */
570 int eflags;
571 /* Where the matching ends. */
572 int match_last;
573 int last_node;
574 /* The state log used by the matcher. */
575 re_dfastate_t **state_log;
576 int state_log_top;
577 /* Back reference cache. */
578 int nbkref_ents;
579 int abkref_ents;
580 struct re_backref_cache_entry *bkref_ents;
581 int max_mb_elem_len;
582 int nsub_tops;
583 int asub_tops;
584 re_sub_match_top_t **sub_tops;
585 } re_match_context_t;
587 typedef struct
589 re_dfastate_t **sifted_states;
590 re_dfastate_t **limited_states;
591 int last_node;
592 int last_str_idx;
593 re_node_set limits;
594 } re_sift_context_t;
596 struct re_fail_stack_ent_t
598 int idx;
599 int node;
600 regmatch_t *regs;
601 re_node_set eps_via_nodes;
604 struct re_fail_stack_t
606 int num;
607 int alloc;
608 struct re_fail_stack_ent_t *stack;
611 struct re_dfa_t
613 re_token_t *nodes;
614 size_t nodes_alloc;
615 size_t nodes_len;
616 int *nexts;
617 int *org_indices;
618 re_node_set *edests;
619 re_node_set *eclosures;
620 re_node_set *inveclosures;
621 struct re_state_table_entry *state_table;
622 re_dfastate_t *init_state;
623 re_dfastate_t *init_state_word;
624 re_dfastate_t *init_state_nl;
625 re_dfastate_t *init_state_begbuf;
626 bin_tree_t *str_tree;
627 bin_tree_storage_t *str_tree_storage;
628 re_bitset_ptr_t sb_char;
629 int str_tree_storage_idx;
631 /* number of subexpressions `re_nsub' is in regex_t. */
632 unsigned int state_hash_mask;
633 int init_node;
634 int nbackref; /* The number of backreference in this dfa. */
636 /* Bitmap expressing which backreference is used. */
637 bitset_word_t used_bkref_map;
638 bitset_word_t completed_bkref_map;
640 unsigned int has_plural_match : 1;
641 /* If this dfa has "multibyte node", which is a backreference or
642 a node which can accept multibyte character or multi character
643 collating element. */
644 unsigned int has_mb_node : 1;
645 unsigned int is_utf8 : 1;
646 unsigned int map_notascii : 1;
647 unsigned int word_ops_used : 1;
648 int mb_cur_max;
649 bitset_t word_char;
650 reg_syntax_t syntax;
651 int *subexp_map;
652 #ifdef DEBUG
653 char* re_str;
654 #endif
655 __libc_lock_define (, lock)
658 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
659 #define re_node_set_remove(set,id) \
660 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
661 #define re_node_set_empty(p) ((p)->nelem = 0)
662 #define re_node_set_free(set) re_free ((set)->elems)
665 typedef enum
667 SB_CHAR,
668 MB_CHAR,
669 EQUIV_CLASS,
670 COLL_SYM,
671 CHAR_CLASS
672 } bracket_elem_type;
674 typedef struct
676 bracket_elem_type type;
677 union
679 unsigned char ch;
680 unsigned char *name;
681 wchar_t wch;
682 } opr;
683 } bracket_elem_t;
686 /* Inline functions for bitset operation. */
687 static void __attribute__ ((unused))
688 bitset_not (bitset_t set)
690 int bitset_i;
691 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
692 set[bitset_i] = ~set[bitset_i];
695 static void __attribute__ ((unused))
696 bitset_merge (bitset_t dest, const bitset_t src)
698 int bitset_i;
699 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
700 dest[bitset_i] |= src[bitset_i];
703 static void __attribute__ ((unused))
704 bitset_mask (bitset_t dest, const bitset_t src)
706 int bitset_i;
707 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
708 dest[bitset_i] &= src[bitset_i];
711 #ifdef RE_ENABLE_I18N
712 /* Inline functions for re_string. */
713 static int
714 internal_function __attribute__ ((pure, unused))
715 re_string_char_size_at (const re_string_t *pstr, int idx)
717 int byte_idx;
718 if (pstr->mb_cur_max == 1)
719 return 1;
720 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
721 if (pstr->wcs[idx + byte_idx] != WEOF)
722 break;
723 return byte_idx;
726 static wint_t
727 internal_function __attribute__ ((pure, unused))
728 re_string_wchar_at (const re_string_t *pstr, int idx)
730 if (pstr->mb_cur_max == 1)
731 return (wint_t) pstr->mbs[idx];
732 return (wint_t) pstr->wcs[idx];
735 # ifndef NOT_IN_libc
736 static int
737 internal_function __attribute__ ((pure, unused))
738 re_string_elem_size_at (const re_string_t *pstr, int idx)
740 # ifdef _LIBC
741 const unsigned char *p, *extra;
742 const int32_t *table, *indirect;
743 # include <locale/weight.h>
744 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
746 if (nrules != 0)
748 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
749 extra = (const unsigned char *)
750 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
751 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
752 _NL_COLLATE_INDIRECTMB);
753 p = pstr->mbs + idx;
754 findidx (&p, pstr->len - idx);
755 return p - pstr->mbs - idx;
757 else
758 # endif /* _LIBC */
759 return 1;
761 # endif
762 #endif /* RE_ENABLE_I18N */
764 #endif /* _REGEX_INTERNAL_H */