regex_internal.h revision 251436 
1/* Extended regular expression matching and search library.
2   Copyright (C) 2002-2012 Free Software Foundation, Inc.
3   This file is part of the GNU C Library.
4   Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
5
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.
10
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.
15
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/>.  */
19
20#ifndef _REGEX_INTERNAL_H
21#define _REGEX_INTERNAL_H 1
22
23#include <assert.h>
24#include <ctype.h>
25#include <stdio.h>
26#include <stdlib.h>
27#include <string.h>
28
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
55
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
60
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
69
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
81
82#ifndef gettext_noop
83/* This define is so xgettext can find the internationalizable
84   strings.  */
85# define gettext_noop(String) String
86#endif
87
88/* For loser systems without the definition.  */
89#ifndef SIZE_MAX
90# define SIZE_MAX ((size_t) -1)
91#endif
92
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
96
97#if __GNUC__ >= 3
98# define BE(expr, val) __builtin_expect (expr, val)
99#else
100# define BE(expr, val) (expr)
101# define inline
102#endif
103
104/* Number of single byte character.  */
105#define SBC_MAX 256
106
107#define COLL_ELEM_LEN_MAX 8
108
109/* The character which represents newline.  */
110#define NEWLINE_CHAR '\n'
111#define WIDE_NEWLINE_CHAR L'\n'
112
113/* Rename to standard API for using out of glibc.  */
114#ifndef _LIBC
115# define __wctype wctype
116# define __iswctype iswctype
117# define __btowc btowc
118# define __mbrtowc mbrtowc
119# define __mempcpy mempcpy
120# define __wcrtomb wcrtomb
121# define __regfree regfree
122# define attribute_hidden
123#endif /* not _LIBC */
124
125#ifdef __GNUC__
126# define __attribute(arg) __attribute__ (arg)
127#else
128# define __attribute(arg)
129#endif
130
131extern const char __re_error_msgid[] attribute_hidden;
132extern const size_t __re_error_msgid_idx[] attribute_hidden;
133
134/* An integer used to represent a set of bits.  It must be unsigned,
135   and must be at least as wide as unsigned int.  */
136typedef unsigned long int bitset_word_t;
137/* All bits set in a bitset_word_t.  */
138#define BITSET_WORD_MAX ULONG_MAX
139/* Number of bits in a bitset_word_t.  */
140#define BITSET_WORD_BITS (sizeof (bitset_word_t) * CHAR_BIT)
141/* Number of bitset_word_t in a bit_set.  */
142#define BITSET_WORDS (SBC_MAX / BITSET_WORD_BITS)
143typedef bitset_word_t bitset_t[BITSET_WORDS];
144typedef bitset_word_t *re_bitset_ptr_t;
145typedef const bitset_word_t *re_const_bitset_ptr_t;
146
147#define bitset_set(set,i) \
148  (set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS)
149#define bitset_clear(set,i) \
150  (set[i / BITSET_WORD_BITS] &= ~((bitset_word_t) 1 << i % BITSET_WORD_BITS))
151#define bitset_contain(set,i) \
152  (set[i / BITSET_WORD_BITS] & ((bitset_word_t) 1 << i % BITSET_WORD_BITS))
153#define bitset_empty(set) memset (set, '\0', sizeof (bitset_t))
154#define bitset_set_all(set) memset (set, '\xff', sizeof (bitset_t))
155#define bitset_copy(dest,src) memcpy (dest, src, sizeof (bitset_t))
156
157#define PREV_WORD_CONSTRAINT 0x0001
158#define PREV_NOTWORD_CONSTRAINT 0x0002
159#define NEXT_WORD_CONSTRAINT 0x0004
160#define NEXT_NOTWORD_CONSTRAINT 0x0008
161#define PREV_NEWLINE_CONSTRAINT 0x0010
162#define NEXT_NEWLINE_CONSTRAINT 0x0020
163#define PREV_BEGBUF_CONSTRAINT 0x0040
164#define NEXT_ENDBUF_CONSTRAINT 0x0080
165#define WORD_DELIM_CONSTRAINT 0x0100
166#define NOT_WORD_DELIM_CONSTRAINT 0x0200
167
168typedef enum
169{
170  INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
171  WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
172  WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
173  INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
174  LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
175  LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
176  BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
177  BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
178  WORD_DELIM = WORD_DELIM_CONSTRAINT,
179  NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
180} re_context_type;
181
182typedef struct
183{
184  int alloc;
185  int nelem;
186  int *elems;
187} re_node_set;
188
189typedef enum
190{
191  NON_TYPE = 0,
192
193  /* Node type, These are used by token, node, tree.  */
194  CHARACTER = 1,
195  END_OF_RE = 2,
196  SIMPLE_BRACKET = 3,
197  OP_BACK_REF = 4,
198  OP_PERIOD = 5,
199#ifdef RE_ENABLE_I18N
200  COMPLEX_BRACKET = 6,
201  OP_UTF8_PERIOD = 7,
202#endif /* RE_ENABLE_I18N */
203
204  /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
205     when the debugger shows values of this enum type.  */
206#define EPSILON_BIT 8
207  OP_OPEN_SUBEXP = EPSILON_BIT | 0,
208  OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
209  OP_ALT = EPSILON_BIT | 2,
210  OP_DUP_ASTERISK = EPSILON_BIT | 3,
211  ANCHOR = EPSILON_BIT | 4,
212
213  /* Tree type, these are used only by tree. */
214  CONCAT = 16,
215  SUBEXP = 17,
216
217  /* Token type, these are used only by token.  */
218  OP_DUP_PLUS = 18,
219  OP_DUP_QUESTION,
220  OP_OPEN_BRACKET,
221  OP_CLOSE_BRACKET,
222  OP_CHARSET_RANGE,
223  OP_OPEN_DUP_NUM,
224  OP_CLOSE_DUP_NUM,
225  OP_NON_MATCH_LIST,
226  OP_OPEN_COLL_ELEM,
227  OP_CLOSE_COLL_ELEM,
228  OP_OPEN_EQUIV_CLASS,
229  OP_CLOSE_EQUIV_CLASS,
230  OP_OPEN_CHAR_CLASS,
231  OP_CLOSE_CHAR_CLASS,
232  OP_WORD,
233  OP_NOTWORD,
234  OP_SPACE,
235  OP_NOTSPACE,
236  BACK_SLASH
237
238} re_token_type_t;
239
240#ifdef RE_ENABLE_I18N
241typedef struct
242{
243  /* Multibyte characters.  */
244  wchar_t *mbchars;
245
246  /* Collating symbols.  */
247# ifdef _LIBC
248  int32_t *coll_syms;
249# endif
250
251  /* Equivalence classes. */
252# ifdef _LIBC
253  int32_t *equiv_classes;
254# endif
255
256  /* Range expressions. */
257# ifdef _LIBC
258  uint32_t *range_starts;
259  uint32_t *range_ends;
260# else /* not _LIBC */
261  wchar_t *range_starts;
262  wchar_t *range_ends;
263# endif /* not _LIBC */
264
265  /* Character classes. */
266  wctype_t *char_classes;
267
268  /* If this character set is the non-matching list.  */
269  unsigned int non_match : 1;
270
271  /* # of multibyte characters.  */
272  int nmbchars;
273
274  /* # of collating symbols.  */
275  int ncoll_syms;
276
277  /* # of equivalence classes. */
278  int nequiv_classes;
279
280  /* # of range expressions. */
281  int nranges;
282
283  /* # of character classes. */
284  int nchar_classes;
285} re_charset_t;
286#endif /* RE_ENABLE_I18N */
287
288typedef struct
289{
290  union
291  {
292    unsigned char c;		/* for CHARACTER */
293    re_bitset_ptr_t sbcset;	/* for SIMPLE_BRACKET */
294#ifdef RE_ENABLE_I18N
295    re_charset_t *mbcset;	/* for COMPLEX_BRACKET */
296#endif /* RE_ENABLE_I18N */
297    int idx;			/* for BACK_REF */
298    re_context_type ctx_type;	/* for ANCHOR */
299  } opr;
300#if __GNUC__ >= 2
301  re_token_type_t type : 8;
302#else
303  re_token_type_t type;
304#endif
305  unsigned int constraint : 10;	/* context constraint */
306  unsigned int duplicated : 1;
307  unsigned int opt_subexp : 1;
308#ifdef RE_ENABLE_I18N
309  unsigned int accept_mb : 1;
310  /* These 2 bits can be moved into the union if needed (e.g. if running out
311     of bits; move opr.c to opr.c.c and move the flags to opr.c.flags).  */
312  unsigned int mb_partial : 1;
313#endif
314  unsigned int word_char : 1;
315} re_token_t;
316
317#define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
318
319struct re_string_t
320{
321  /* Indicate the raw buffer which is the original string passed as an
322     argument of regexec(), re_search(), etc..  */
323  const unsigned char *raw_mbs;
324  /* Store the multibyte string.  In case of "case insensitive mode" like
325     REG_ICASE, upper cases of the string are stored, otherwise MBS points
326     the same address that RAW_MBS points.  */
327  unsigned char *mbs;
328#ifdef RE_ENABLE_I18N
329  /* Store the wide character string which is corresponding to MBS.  */
330  wint_t *wcs;
331  int *offsets;
332  mbstate_t cur_state;
333#endif
334  /* Index in RAW_MBS.  Each character mbs[i] corresponds to
335     raw_mbs[raw_mbs_idx + i].  */
336  int raw_mbs_idx;
337  /* The length of the valid characters in the buffers.  */
338  int valid_len;
339  /* The corresponding number of bytes in raw_mbs array.  */
340  int valid_raw_len;
341  /* The length of the buffers MBS and WCS.  */
342  int bufs_len;
343  /* The index in MBS, which is updated by re_string_fetch_byte.  */
344  int cur_idx;
345  /* length of RAW_MBS array.  */
346  int raw_len;
347  /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN.  */
348  int len;
349  /* End of the buffer may be shorter than its length in the cases such
350     as re_match_2, re_search_2.  Then, we use STOP for end of the buffer
351     instead of LEN.  */
352  int raw_stop;
353  /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS.  */
354  int stop;
355
356  /* The context of mbs[0].  We store the context independently, since
357     the context of mbs[0] may be different from raw_mbs[0], which is
358     the beginning of the input string.  */
359  unsigned int tip_context;
360  /* The translation passed as a part of an argument of re_compile_pattern.  */
361  RE_TRANSLATE_TYPE trans;
362  /* Copy of re_dfa_t's word_char.  */
363  re_const_bitset_ptr_t word_char;
364  /* 1 if REG_ICASE.  */
365  unsigned char icase;
366  unsigned char is_utf8;
367  unsigned char map_notascii;
368  unsigned char mbs_allocated;
369  unsigned char offsets_needed;
370  unsigned char newline_anchor;
371  unsigned char word_ops_used;
372  int mb_cur_max;
373};
374typedef struct re_string_t re_string_t;
375
376
377struct re_dfa_t;
378typedef struct re_dfa_t re_dfa_t;
379
380#ifndef _LIBC
381# ifdef __i386__
382#  define internal_function   __attribute ((regparm (3), stdcall))
383# else
384#  define internal_function
385# endif
386#endif
387
388#ifndef NOT_IN_libc
389static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
390						int new_buf_len)
391     internal_function;
392# ifdef RE_ENABLE_I18N
393static void build_wcs_buffer (re_string_t *pstr) internal_function;
394static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
395  internal_function;
396# endif /* RE_ENABLE_I18N */
397static void build_upper_buffer (re_string_t *pstr) internal_function;
398static void re_string_translate_buffer (re_string_t *pstr) internal_function;
399static unsigned int re_string_context_at (const re_string_t *input, int idx,
400					  int eflags)
401     internal_function __attribute ((pure));
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))
419
420#include <alloca.h>
421
422#ifndef _LIBC
423# if HAVE_ALLOCA
424/* The OS usually guarantees only one guard page at the bottom of the stack,
425   and a page size can be as small as 4096 bytes.  So we cannot safely
426   allocate anything larger than 4096 bytes.  Also care for the possibility
427   of a few compiler-allocated temporary stack slots.  */
428#  define __libc_use_alloca(n) ((n) < 4032)
429# else
430/* alloca is implemented with malloc, so just use malloc.  */
431#  define __libc_use_alloca(n) 0
432# endif
433#endif
434
435#define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
436#define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
437#define re_free(p) free (p)
438
439struct bin_tree_t
440{
441  struct bin_tree_t *parent;
442  struct bin_tree_t *left;
443  struct bin_tree_t *right;
444  struct bin_tree_t *first;
445  struct bin_tree_t *next;
446
447  re_token_t token;
448
449  /* `node_idx' is the index in dfa->nodes, if `type' == 0.
450     Otherwise `type' indicate the type of this node.  */
451  int node_idx;
452};
453typedef struct bin_tree_t bin_tree_t;
454
455#define BIN_TREE_STORAGE_SIZE \
456  ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
457
458struct bin_tree_storage_t
459{
460  struct bin_tree_storage_t *next;
461  bin_tree_t data[BIN_TREE_STORAGE_SIZE];
462};
463typedef struct bin_tree_storage_t bin_tree_storage_t;
464
465#define CONTEXT_WORD 1
466#define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
467#define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
468#define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
469
470#define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
471#define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
472#define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
473#define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
474#define IS_ORDINARY_CONTEXT(c) ((c) == 0)
475
476#define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
477#define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
478#define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
479#define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
480
481#define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
482 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
483  || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
484  || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
485  || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
486
487#define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
488 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
489  || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
490  || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
491  || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
492
493struct re_dfastate_t
494{
495  unsigned int hash;
496  re_node_set nodes;
497  re_node_set non_eps_nodes;
498  re_node_set inveclosure;
499  re_node_set *entrance_nodes;
500  struct re_dfastate_t **trtable, **word_trtable;
501  unsigned int context : 4;
502  unsigned int halt : 1;
503  /* If this state can accept `multi byte'.
504     Note that we refer to multibyte characters, and multi character
505     collating elements as `multi byte'.  */
506  unsigned int accept_mb : 1;
507  /* If this state has backreference node(s).  */
508  unsigned int has_backref : 1;
509  unsigned int has_constraint : 1;
510};
511typedef struct re_dfastate_t re_dfastate_t;
512
513struct re_state_table_entry
514{
515  int num;
516  int alloc;
517  re_dfastate_t **array;
518};
519
520/* Array type used in re_sub_match_last_t and re_sub_match_top_t.  */
521
522typedef struct
523{
524  int next_idx;
525  int alloc;
526  re_dfastate_t **array;
527} state_array_t;
528
529/* Store information about the node NODE whose type is OP_CLOSE_SUBEXP.  */
530
531typedef struct
532{
533  int node;
534  int str_idx; /* The position NODE match at.  */
535  state_array_t path;
536} re_sub_match_last_t;
537
538/* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
539   And information about the node, whose type is OP_CLOSE_SUBEXP,
540   corresponding to NODE is stored in LASTS.  */
541
542typedef struct
543{
544  int str_idx;
545  int node;
546  state_array_t *path;
547  int alasts; /* Allocation size of LASTS.  */
548  int nlasts; /* The number of LASTS.  */
549  re_sub_match_last_t **lasts;
550} re_sub_match_top_t;
551
552struct re_backref_cache_entry
553{
554  int node;
555  int str_idx;
556  int subexp_from;
557  int subexp_to;
558  char more;
559  char unused;
560  unsigned short int eps_reachable_subexps_map;
561};
562
563typedef struct
564{
565  /* The string object corresponding to the input string.  */
566  re_string_t input;
567#if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
568  const re_dfa_t *const dfa;
569#else
570  const re_dfa_t *dfa;
571#endif
572  /* EFLAGS of the argument of regexec.  */
573  int eflags;
574  /* Where the matching ends.  */
575  int match_last;
576  int last_node;
577  /* The state log used by the matcher.  */
578  re_dfastate_t **state_log;
579  int state_log_top;
580  /* Back reference cache.  */
581  int nbkref_ents;
582  int abkref_ents;
583  struct re_backref_cache_entry *bkref_ents;
584  int max_mb_elem_len;
585  int nsub_tops;
586  int asub_tops;
587  re_sub_match_top_t **sub_tops;
588} re_match_context_t;
589
590typedef struct
591{
592  re_dfastate_t **sifted_states;
593  re_dfastate_t **limited_states;
594  int last_node;
595  int last_str_idx;
596  re_node_set limits;
597} re_sift_context_t;
598
599struct re_fail_stack_ent_t
600{
601  int idx;
602  int node;
603  regmatch_t *regs;
604  re_node_set eps_via_nodes;
605};
606
607struct re_fail_stack_t
608{
609  int num;
610  int alloc;
611  struct re_fail_stack_ent_t *stack;
612};
613
614struct re_dfa_t
615{
616  re_token_t *nodes;
617  size_t nodes_alloc;
618  size_t nodes_len;
619  int *nexts;
620  int *org_indices;
621  re_node_set *edests;
622  re_node_set *eclosures;
623  re_node_set *inveclosures;
624  struct re_state_table_entry *state_table;
625  re_dfastate_t *init_state;
626  re_dfastate_t *init_state_word;
627  re_dfastate_t *init_state_nl;
628  re_dfastate_t *init_state_begbuf;
629  bin_tree_t *str_tree;
630  bin_tree_storage_t *str_tree_storage;
631  re_bitset_ptr_t sb_char;
632  int str_tree_storage_idx;
633
634  /* number of subexpressions `re_nsub' is in regex_t.  */
635  unsigned int state_hash_mask;
636  int init_node;
637  int nbackref; /* The number of backreference in this dfa.  */
638
639  /* Bitmap expressing which backreference is used.  */
640  bitset_word_t used_bkref_map;
641  bitset_word_t completed_bkref_map;
642
643  unsigned int has_plural_match : 1;
644  /* If this dfa has "multibyte node", which is a backreference or
645     a node which can accept multibyte character or multi character
646     collating element.  */
647  unsigned int has_mb_node : 1;
648  unsigned int is_utf8 : 1;
649  unsigned int map_notascii : 1;
650  unsigned int word_ops_used : 1;
651  int mb_cur_max;
652  bitset_t word_char;
653  reg_syntax_t syntax;
654  int *subexp_map;
655#ifdef DEBUG
656  char* re_str;
657#endif
658  __libc_lock_define (, lock)
659};
660
661#define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
662#define re_node_set_remove(set,id) \
663  (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
664#define re_node_set_empty(p) ((p)->nelem = 0)
665#define re_node_set_free(set) re_free ((set)->elems)
666
667
668typedef enum
669{
670  SB_CHAR,
671  MB_CHAR,
672  EQUIV_CLASS,
673  COLL_SYM,
674  CHAR_CLASS
675} bracket_elem_type;
676
677typedef struct
678{
679  bracket_elem_type type;
680  union
681  {
682    unsigned char ch;
683    unsigned char *name;
684    wchar_t wch;
685  } opr;
686} bracket_elem_t;
687
688
689/* Inline functions for bitset operation.  */
690static inline void
691bitset_not (bitset_t set)
692{
693  int bitset_i;
694  for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
695    set[bitset_i] = ~set[bitset_i];
696}
697
698static inline void
699bitset_merge (bitset_t dest, const bitset_t src)
700{
701  int bitset_i;
702  for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
703    dest[bitset_i] |= src[bitset_i];
704}
705
706static inline void
707bitset_mask (bitset_t dest, const bitset_t src)
708{
709  int bitset_i;
710  for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
711    dest[bitset_i] &= src[bitset_i];
712}
713
714#ifdef RE_ENABLE_I18N
715/* Inline functions for re_string.  */
716static inline int
717internal_function __attribute ((pure))
718re_string_char_size_at (const re_string_t *pstr, int idx)
719{
720  int byte_idx;
721  if (pstr->mb_cur_max == 1)
722    return 1;
723  for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
724    if (pstr->wcs[idx + byte_idx] != WEOF)
725      break;
726  return byte_idx;
727}
728
729static inline wint_t
730internal_function __attribute ((pure))
731re_string_wchar_at (const re_string_t *pstr, int idx)
732{
733  if (pstr->mb_cur_max == 1)
734    return (wint_t) pstr->mbs[idx];
735  return (wint_t) pstr->wcs[idx];
736}
737
738# ifndef NOT_IN_libc
739static int
740internal_function __attribute ((pure))
741re_string_elem_size_at (const re_string_t *pstr, int idx)
742{
743#  ifdef _LIBC
744  const unsigned char *p, *extra;
745  const int32_t *table, *indirect;
746#   include <locale/weight.h>
747  uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
748
749  if (nrules != 0)
750    {
751      table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
752      extra = (const unsigned char *)
753	_NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
754      indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
755						_NL_COLLATE_INDIRECTMB);
756      p = pstr->mbs + idx;
757      findidx (&p, pstr->len - idx);
758      return p - pstr->mbs - idx;
759    }
760  else
761#  endif /* _LIBC */
762    return 1;
763}
764# endif
765#endif /* RE_ENABLE_I18N */
766
767#endif /*  _REGEX_INTERNAL_H */
768