Updated to fedora-glibc-20041110T0839
[glibc.git] / posix / regex_internal.h
blob023056c0282793c10f68e5ed1d373c8dd59d07bd
1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002, 2003, 2004 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, write to the Free
18 Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
19 02111-1307 USA. */
21 #ifndef _REGEX_INTERNAL_H
22 #define _REGEX_INTERNAL_H 1
24 #include <assert.h>
25 #include <ctype.h>
26 #include <stdio.h>
27 #include <stdlib.h>
28 #include <string.h>
30 #if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET || defined _LIBC
31 # include <langinfo.h>
32 #endif
33 #if defined HAVE_LOCALE_H || defined _LIBC
34 # include <locale.h>
35 #endif
36 #if defined HAVE_WCHAR_H || defined _LIBC
37 # include <wchar.h>
38 #endif /* HAVE_WCHAR_H || _LIBC */
39 #if defined HAVE_WCTYPE_H || defined _LIBC
40 # include <wctype.h>
41 #endif /* HAVE_WCTYPE_H || _LIBC */
43 /* In case that the system doesn't have isblank(). */
44 #if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank
45 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
46 #endif
48 #ifdef _LIBC
49 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
50 # define _RE_DEFINE_LOCALE_FUNCTIONS 1
51 # include <locale/localeinfo.h>
52 # include <locale/elem-hash.h>
53 # include <locale/coll-lookup.h>
54 # endif
55 #endif
57 /* This is for other GNU distributions with internationalized messages. */
58 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
59 # include <libintl.h>
60 # ifdef _LIBC
61 # undef gettext
62 # define gettext(msgid) \
63 INTUSE(__dcgettext) (INTUSE(_libc_intl_domainname), msgid, LC_MESSAGES)
64 # endif
65 #else
66 # define gettext(msgid) (msgid)
67 #endif
69 #ifndef gettext_noop
70 /* This define is so xgettext can find the internationalizable
71 strings. */
72 # define gettext_noop(String) String
73 #endif
75 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC
76 # define RE_ENABLE_I18N
77 #endif
79 #if __GNUC__ >= 3
80 # define BE(expr, val) __builtin_expect (expr, val)
81 #else
82 # define BE(expr, val) (expr)
83 # define inline
84 #endif
86 /* Number of bits in a byte. */
87 #define BYTE_BITS 8
88 /* Number of single byte character. */
89 #define SBC_MAX 256
91 #define COLL_ELEM_LEN_MAX 8
93 /* The character which represents newline. */
94 #define NEWLINE_CHAR '\n'
95 #define WIDE_NEWLINE_CHAR L'\n'
97 /* Rename to standard API for using out of glibc. */
98 #ifndef _LIBC
99 # define __wctype wctype
100 # define __iswctype iswctype
101 # define __btowc btowc
102 # define __mempcpy mempcpy
103 # define __wcrtomb wcrtomb
104 # define __regfree regfree
105 # define attribute_hidden
106 #endif /* not _LIBC */
108 #ifdef __GNUC__
109 # define __attribute(arg) __attribute__ (arg)
110 #else
111 # define __attribute(arg)
112 #endif
114 extern const char __re_error_msgid[] attribute_hidden;
115 extern const size_t __re_error_msgid_idx[] attribute_hidden;
117 /* Number of bits in an unsinged int. */
118 #define UINT_BITS (sizeof (unsigned int) * BYTE_BITS)
119 /* Number of unsigned int in an bit_set. */
120 #define BITSET_UINTS ((SBC_MAX + UINT_BITS - 1) / UINT_BITS)
121 typedef unsigned int bitset[BITSET_UINTS];
122 typedef unsigned int *re_bitset_ptr_t;
123 typedef const unsigned int *re_const_bitset_ptr_t;
125 #define bitset_set(set,i) (set[i / UINT_BITS] |= 1 << i % UINT_BITS)
126 #define bitset_clear(set,i) (set[i / UINT_BITS] &= ~(1 << i % UINT_BITS))
127 #define bitset_contain(set,i) (set[i / UINT_BITS] & (1 << i % UINT_BITS))
128 #define bitset_empty(set) memset (set, 0, sizeof (unsigned int) * BITSET_UINTS)
129 #define bitset_set_all(set) \
130 memset (set, 255, sizeof (unsigned int) * BITSET_UINTS)
131 #define bitset_copy(dest,src) \
132 memcpy (dest, src, sizeof (unsigned int) * BITSET_UINTS)
133 static inline void bitset_not (bitset set);
134 static inline void bitset_merge (bitset dest, const bitset src);
135 static inline void bitset_not_merge (bitset dest, const bitset src);
136 static inline void bitset_mask (bitset dest, const bitset src);
138 #define PREV_WORD_CONSTRAINT 0x0001
139 #define PREV_NOTWORD_CONSTRAINT 0x0002
140 #define NEXT_WORD_CONSTRAINT 0x0004
141 #define NEXT_NOTWORD_CONSTRAINT 0x0008
142 #define PREV_NEWLINE_CONSTRAINT 0x0010
143 #define NEXT_NEWLINE_CONSTRAINT 0x0020
144 #define PREV_BEGBUF_CONSTRAINT 0x0040
145 #define NEXT_ENDBUF_CONSTRAINT 0x0080
146 #define DUMMY_CONSTRAINT 0x0100
148 typedef enum
150 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
151 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
152 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
153 LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
154 LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
155 BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
156 BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
157 WORD_DELIM = DUMMY_CONSTRAINT
158 } re_context_type;
160 typedef struct
162 int alloc;
163 int nelem;
164 int *elems;
165 } re_node_set;
167 typedef enum
169 NON_TYPE = 0,
171 /* Node type, These are used by token, node, tree. */
172 CHARACTER = 1,
173 END_OF_RE = 2,
174 SIMPLE_BRACKET = 3,
175 OP_BACK_REF = 4,
176 OP_PERIOD = 5,
177 #ifdef RE_ENABLE_I18N
178 COMPLEX_BRACKET = 6,
179 OP_UTF8_PERIOD = 7,
180 #endif /* RE_ENABLE_I18N */
182 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
183 when the debugger shows values of this enum type. */
184 #define EPSILON_BIT 8
185 OP_OPEN_SUBEXP = EPSILON_BIT | 0,
186 OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
187 OP_ALT = EPSILON_BIT | 2,
188 OP_DUP_ASTERISK = EPSILON_BIT | 3,
189 OP_DUP_PLUS = EPSILON_BIT | 4,
190 OP_DUP_QUESTION = EPSILON_BIT | 5,
191 ANCHOR = EPSILON_BIT | 6,
193 /* Tree type, these are used only by tree. */
194 CONCAT = 16,
196 /* Token type, these are used only by token. */
197 OP_OPEN_BRACKET = 17,
198 OP_CLOSE_BRACKET,
199 OP_CHARSET_RANGE,
200 OP_OPEN_DUP_NUM,
201 OP_CLOSE_DUP_NUM,
202 OP_NON_MATCH_LIST,
203 OP_OPEN_COLL_ELEM,
204 OP_CLOSE_COLL_ELEM,
205 OP_OPEN_EQUIV_CLASS,
206 OP_CLOSE_EQUIV_CLASS,
207 OP_OPEN_CHAR_CLASS,
208 OP_CLOSE_CHAR_CLASS,
209 OP_WORD,
210 OP_NOTWORD,
211 OP_SPACE,
212 OP_NOTSPACE,
213 BACK_SLASH
215 } re_token_type_t;
217 #ifdef RE_ENABLE_I18N
218 typedef struct
220 /* Multibyte characters. */
221 wchar_t *mbchars;
223 /* Collating symbols. */
224 # ifdef _LIBC
225 int32_t *coll_syms;
226 # endif
228 /* Equivalence classes. */
229 # ifdef _LIBC
230 int32_t *equiv_classes;
231 # endif
233 /* Range expressions. */
234 # ifdef _LIBC
235 uint32_t *range_starts;
236 uint32_t *range_ends;
237 # else /* not _LIBC */
238 wchar_t *range_starts;
239 wchar_t *range_ends;
240 # endif /* not _LIBC */
242 /* Character classes. */
243 wctype_t *char_classes;
245 /* If this character set is the non-matching list. */
246 unsigned int non_match : 1;
248 /* # of multibyte characters. */
249 int nmbchars;
251 /* # of collating symbols. */
252 int ncoll_syms;
254 /* # of equivalence classes. */
255 int nequiv_classes;
257 /* # of range expressions. */
258 int nranges;
260 /* # of character classes. */
261 int nchar_classes;
262 } re_charset_t;
263 #endif /* RE_ENABLE_I18N */
265 typedef struct
267 union
269 unsigned char c; /* for CHARACTER */
270 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
271 #ifdef RE_ENABLE_I18N
272 re_charset_t *mbcset; /* for COMPLEX_BRACKET */
273 #endif /* RE_ENABLE_I18N */
274 int idx; /* for BACK_REF */
275 re_context_type ctx_type; /* for ANCHOR */
276 } opr;
277 #if __GNUC__ >= 2
278 re_token_type_t type : 8;
279 #else
280 re_token_type_t type;
281 #endif
282 unsigned int constraint : 10; /* context constraint */
283 unsigned int duplicated : 1;
284 unsigned int opt_subexp : 1;
285 #ifdef RE_ENABLE_I18N
286 /* These 2 bits can be moved into the union if needed (e.g. if running out
287 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
288 unsigned int mb_partial : 1;
289 #endif
290 unsigned int word_char : 1;
291 } re_token_t;
293 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
294 #define ACCEPT_MB_NODE(type) \
295 ((type) >= OP_PERIOD && (type) <= OP_UTF8_PERIOD)
297 struct re_string_t
299 /* Indicate the raw buffer which is the original string passed as an
300 argument of regexec(), re_search(), etc.. */
301 const unsigned char *raw_mbs;
302 /* Store the multibyte string. In case of "case insensitive mode" like
303 REG_ICASE, upper cases of the string are stored, otherwise MBS points
304 the same address that RAW_MBS points. */
305 unsigned char *mbs;
306 #ifdef RE_ENABLE_I18N
307 /* Store the wide character string which is corresponding to MBS. */
308 wint_t *wcs;
309 int *offsets;
310 mbstate_t cur_state;
311 #endif
312 /* Index in RAW_MBS. Each character mbs[i] corresponds to
313 raw_mbs[raw_mbs_idx + i]. */
314 int raw_mbs_idx;
315 /* The length of the valid characters in the buffers. */
316 int valid_len;
317 /* The corresponding number of bytes in raw_mbs array. */
318 int valid_raw_len;
319 /* The length of the buffers MBS and WCS. */
320 int bufs_len;
321 /* The index in MBS, which is updated by re_string_fetch_byte. */
322 int cur_idx;
323 /* length of RAW_MBS array. */
324 int raw_len;
325 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
326 int len;
327 /* End of the buffer may be shorter than its length in the cases such
328 as re_match_2, re_search_2. Then, we use STOP for end of the buffer
329 instead of LEN. */
330 int raw_stop;
331 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
332 int stop;
334 /* The context of mbs[0]. We store the context independently, since
335 the context of mbs[0] may be different from raw_mbs[0], which is
336 the beginning of the input string. */
337 unsigned int tip_context;
338 /* The translation passed as a part of an argument of re_compile_pattern. */
339 unsigned RE_TRANSLATE_TYPE trans;
340 /* Copy of re_dfa_t's word_char. */
341 re_const_bitset_ptr_t word_char;
342 /* 1 if REG_ICASE. */
343 unsigned char icase;
344 unsigned char is_utf8;
345 unsigned char map_notascii;
346 unsigned char mbs_allocated;
347 unsigned char offsets_needed;
348 unsigned char newline_anchor;
349 unsigned char word_ops_used;
350 int mb_cur_max;
352 typedef struct re_string_t re_string_t;
355 struct re_dfa_t;
356 typedef struct re_dfa_t re_dfa_t;
358 #ifndef _LIBC
359 # ifdef __i386__
360 # define internal_function __attribute ((regparm (3), stdcall))
361 # else
362 # define internal_function
363 # endif
364 #endif
366 #ifndef RE_NO_INTERNAL_PROTOTYPES
367 static reg_errcode_t re_string_allocate (re_string_t *pstr, const char *str,
368 int len, int init_len,
369 RE_TRANSLATE_TYPE trans, int icase,
370 const re_dfa_t *dfa)
371 internal_function;
372 static reg_errcode_t re_string_construct (re_string_t *pstr, const char *str,
373 int len, RE_TRANSLATE_TYPE trans,
374 int icase, const re_dfa_t *dfa)
375 internal_function;
376 static reg_errcode_t re_string_reconstruct (re_string_t *pstr, int idx,
377 int eflags) internal_function;
378 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
379 int new_buf_len)
380 internal_function;
381 # ifdef RE_ENABLE_I18N
382 static void build_wcs_buffer (re_string_t *pstr) internal_function;
383 static int build_wcs_upper_buffer (re_string_t *pstr) internal_function;
384 # endif /* RE_ENABLE_I18N */
385 static void build_upper_buffer (re_string_t *pstr) internal_function;
386 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
387 static void re_string_destruct (re_string_t *pstr) internal_function;
388 # ifdef RE_ENABLE_I18N
389 static int re_string_elem_size_at (const re_string_t *pstr, int idx)
390 internal_function;
391 static inline int re_string_char_size_at (const re_string_t *pstr, int idx)
392 internal_function;
393 static inline wint_t re_string_wchar_at (const re_string_t *pstr, int idx)
394 internal_function;
395 # endif /* RE_ENABLE_I18N */
396 static unsigned int re_string_context_at (const re_string_t *input, int idx,
397 int eflags) internal_function;
398 static unsigned char re_string_peek_byte_case (const re_string_t *pstr,
399 int idx) internal_function;
400 static unsigned char re_string_fetch_byte_case (re_string_t *pstr)
401 internal_function;
402 #endif
403 #define re_string_peek_byte(pstr, offset) \
404 ((pstr)->mbs[(pstr)->cur_idx + offset])
405 #define re_string_fetch_byte(pstr) \
406 ((pstr)->mbs[(pstr)->cur_idx++])
407 #define re_string_first_byte(pstr, idx) \
408 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
409 #define re_string_is_single_byte_char(pstr, idx) \
410 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
411 || (pstr)->wcs[(idx) + 1] != WEOF))
412 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
413 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
414 #define re_string_get_buffer(pstr) ((pstr)->mbs)
415 #define re_string_length(pstr) ((pstr)->len)
416 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
417 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
418 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
420 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
421 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
422 #define re_free(p) free (p)
424 struct bin_tree_t
426 struct bin_tree_t *parent;
427 struct bin_tree_t *left;
428 struct bin_tree_t *right;
430 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
431 Otherwise `type' indicate the type of this node. */
432 re_token_type_t type;
433 int node_idx;
435 int first;
436 int next;
437 re_node_set eclosure;
439 typedef struct bin_tree_t bin_tree_t;
441 #define BIN_TREE_STORAGE_SIZE \
442 ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
444 struct bin_tree_storage_t
446 struct bin_tree_storage_t *next;
447 bin_tree_t data[BIN_TREE_STORAGE_SIZE];
449 typedef struct bin_tree_storage_t bin_tree_storage_t;
451 #define CONTEXT_WORD 1
452 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
453 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
454 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
456 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
457 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
458 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
459 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
460 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
462 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
463 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
464 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
465 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
467 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
468 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
469 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
470 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
471 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
473 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
474 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
475 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
476 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
477 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
479 struct re_dfastate_t
481 unsigned int hash;
482 re_node_set nodes;
483 re_node_set *entrance_nodes;
484 struct re_dfastate_t **trtable;
485 unsigned int context : 4;
486 unsigned int halt : 1;
487 /* If this state can accept `multi byte'.
488 Note that we refer to multibyte characters, and multi character
489 collating elements as `multi byte'. */
490 unsigned int accept_mb : 1;
491 /* If this state has backreference node(s). */
492 unsigned int has_backref : 1;
493 unsigned int has_constraint : 1;
494 unsigned int word_trtable : 1;
496 typedef struct re_dfastate_t re_dfastate_t;
498 typedef struct
500 /* start <= node < end */
501 int start;
502 int end;
503 } re_subexp_t;
505 struct re_state_table_entry
507 int num;
508 int alloc;
509 re_dfastate_t **array;
512 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
514 typedef struct
516 int next_idx;
517 int alloc;
518 re_dfastate_t **array;
519 } state_array_t;
521 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
523 typedef struct
525 int node;
526 int str_idx; /* The position NODE match at. */
527 state_array_t path;
528 } re_sub_match_last_t;
530 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
531 And information about the node, whose type is OP_CLOSE_SUBEXP,
532 corresponding to NODE is stored in LASTS. */
534 typedef struct
536 int str_idx;
537 int node;
538 int next_last_offset;
539 state_array_t *path;
540 int alasts; /* Allocation size of LASTS. */
541 int nlasts; /* The number of LASTS. */
542 re_sub_match_last_t **lasts;
543 } re_sub_match_top_t;
545 struct re_backref_cache_entry
547 int node;
548 int str_idx;
549 int subexp_from;
550 int subexp_to;
551 /* We need only one byte from the following field. If other small
552 fields are added the type could be changed to 'char'. */
553 int more;
556 typedef struct
558 /* The string object corresponding to the input string. */
559 re_string_t input;
560 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
561 re_dfa_t *const dfa;
562 #else
563 re_dfa_t *dfa;
564 #endif
565 /* EFLAGS of the argument of regexec. */
566 int eflags;
567 /* Where the matching ends. */
568 int match_last;
569 int last_node;
570 /* The state log used by the matcher. */
571 re_dfastate_t **state_log;
572 int state_log_top;
573 /* Back reference cache. */
574 int nbkref_ents;
575 int abkref_ents;
576 struct re_backref_cache_entry *bkref_ents;
577 int max_mb_elem_len;
578 int nsub_tops;
579 int asub_tops;
580 re_sub_match_top_t **sub_tops;
581 } re_match_context_t;
583 typedef struct
585 re_dfastate_t **sifted_states;
586 re_dfastate_t **limited_states;
587 int last_node;
588 int last_str_idx;
589 re_node_set limits;
590 } re_sift_context_t;
592 struct re_fail_stack_ent_t
594 int idx;
595 int node;
596 regmatch_t *regs;
597 re_node_set eps_via_nodes;
600 struct re_fail_stack_t
602 int num;
603 int alloc;
604 struct re_fail_stack_ent_t *stack;
607 struct re_dfa_t
609 re_subexp_t *subexps;
610 re_token_t *nodes;
611 int nodes_alloc;
612 int nodes_len;
613 int *nexts;
614 int *org_indices;
615 re_node_set *edests;
616 re_node_set *eclosures;
617 re_node_set *inveclosures;
618 struct re_state_table_entry *state_table;
619 re_dfastate_t *init_state;
620 re_dfastate_t *init_state_word;
621 re_dfastate_t *init_state_nl;
622 re_dfastate_t *init_state_begbuf;
623 bin_tree_t *str_tree;
624 bin_tree_storage_t *str_tree_storage;
625 re_bitset_ptr_t sb_char;
626 int str_tree_storage_idx;
628 /* number of subexpressions `re_nsub' is in regex_t. */
629 int subexps_alloc;
630 unsigned int state_hash_mask;
631 int states_alloc;
632 int init_node;
633 int nbackref; /* The number of backreference in this dfa. */
634 /* Bitmap expressing which backreference is used. */
635 unsigned int used_bkref_map;
636 unsigned int has_plural_match : 1;
637 /* If this dfa has "multibyte node", which is a backreference or
638 a node which can accept multibyte character or multi character
639 collating element. */
640 unsigned int has_mb_node : 1;
641 unsigned int is_utf8 : 1;
642 unsigned int map_notascii : 1;
643 unsigned int word_ops_used : 1;
644 int mb_cur_max;
645 bitset word_char;
646 reg_syntax_t syntax;
647 #ifdef DEBUG
648 char* re_str;
649 #endif
652 #ifndef RE_NO_INTERNAL_PROTOTYPES
653 static reg_errcode_t re_node_set_alloc (re_node_set *set, int size) internal_function;
654 static reg_errcode_t re_node_set_init_1 (re_node_set *set, int elem) internal_function;
655 static reg_errcode_t re_node_set_init_2 (re_node_set *set, int elem1,
656 int elem2) internal_function;
657 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
658 static reg_errcode_t re_node_set_init_copy (re_node_set *dest,
659 const re_node_set *src) internal_function;
660 static reg_errcode_t re_node_set_add_intersect (re_node_set *dest,
661 const re_node_set *src1,
662 const re_node_set *src2) internal_function;
663 static reg_errcode_t re_node_set_init_union (re_node_set *dest,
664 const re_node_set *src1,
665 const re_node_set *src2) internal_function;
666 static reg_errcode_t re_node_set_merge (re_node_set *dest,
667 const re_node_set *src) internal_function;
668 static int re_node_set_insert (re_node_set *set, int elem) internal_function;
669 static int re_node_set_compare (const re_node_set *set1,
670 const re_node_set *set2) internal_function;
671 static int re_node_set_contains (const re_node_set *set, int elem) internal_function;
672 static void re_node_set_remove_at (re_node_set *set, int idx) internal_function;
673 #define re_node_set_remove(set,id) \
674 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
675 #define re_node_set_empty(p) ((p)->nelem = 0)
676 #define re_node_set_free(set) re_free ((set)->elems)
677 static int re_dfa_add_node (re_dfa_t *dfa, re_token_t token, int mode) internal_function;
678 static re_dfastate_t *re_acquire_state (reg_errcode_t *err, re_dfa_t *dfa,
679 const re_node_set *nodes) internal_function;
680 static re_dfastate_t *re_acquire_state_context (reg_errcode_t *err,
681 re_dfa_t *dfa,
682 const re_node_set *nodes,
683 unsigned int context) internal_function;
684 static void free_state (re_dfastate_t *state) internal_function;
685 #endif
688 typedef enum
690 SB_CHAR,
691 MB_CHAR,
692 EQUIV_CLASS,
693 COLL_SYM,
694 CHAR_CLASS
695 } bracket_elem_type;
697 typedef struct
699 bracket_elem_type type;
700 union
702 unsigned char ch;
703 unsigned char *name;
704 wchar_t wch;
705 } opr;
706 } bracket_elem_t;
709 /* Inline functions for bitset operation. */
710 static inline void
711 bitset_not (bitset set)
713 int bitset_i;
714 for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i)
715 set[bitset_i] = ~set[bitset_i];
718 static inline void
719 bitset_merge (bitset dest, const bitset src)
721 int bitset_i;
722 for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i)
723 dest[bitset_i] |= src[bitset_i];
726 static inline void
727 bitset_not_merge (bitset dest, const bitset src)
729 int i;
730 for (i = 0; i < BITSET_UINTS; ++i)
731 dest[i] |= ~src[i];
734 static inline void
735 bitset_mask (bitset dest, const bitset src)
737 int bitset_i;
738 for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i)
739 dest[bitset_i] &= src[bitset_i];
742 #if defined RE_ENABLE_I18N && !defined RE_NO_INTERNAL_PROTOTYPES
743 /* Inline functions for re_string. */
744 static inline int
745 internal_function
746 re_string_char_size_at (const re_string_t *pstr, int idx)
748 int byte_idx;
749 if (pstr->mb_cur_max == 1)
750 return 1;
751 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
752 if (pstr->wcs[idx + byte_idx] != WEOF)
753 break;
754 return byte_idx;
757 static inline wint_t
758 internal_function
759 re_string_wchar_at (const re_string_t *pstr, int idx)
761 if (pstr->mb_cur_max == 1)
762 return (wint_t) pstr->mbs[idx];
763 return (wint_t) pstr->wcs[idx];
766 static int
767 internal_function
768 re_string_elem_size_at (const re_string_t *pstr, int idx)
770 #ifdef _LIBC
771 const unsigned char *p, *extra;
772 const int32_t *table, *indirect;
773 int32_t tmp;
774 # include <locale/weight.h>
775 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
777 if (nrules != 0)
779 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
780 extra = (const unsigned char *)
781 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
782 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
783 _NL_COLLATE_INDIRECTMB);
784 p = pstr->mbs + idx;
785 tmp = findidx (&p);
786 return p - pstr->mbs - idx;
788 else
789 #endif /* _LIBC */
790 return 1;
792 #endif /* RE_ENABLE_I18N */
794 #endif /* _REGEX_INTERNAL_H */