libc/misc/regex/regex_internal.h

   1 /* Extended regular expression matching and search library.
   2    Copyright (C) 2002, 2003, 2004, 2005 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 <stdlib.h>
  26 #include <string.h>
  27
  28 #if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET
  29 # include <langinfo.h>
  30 #endif
  31 #if defined HAVE_LOCALE_H
  32 # include <locale.h>
  33 #endif
  34 #if defined HAVE_WCHAR_H
  35 # include <wchar.h>
  36 #endif
  37 #if defined HAVE_WCTYPE_H
  38 # include <wctype.h>
  39 #endif
  40 #if defined HAVE_STDBOOL_H
  41 # include <stdbool.h>
  42 #endif
  43 #if defined HAVE_STDINT_H
  44 # include <stdint.h>
  45 #endif
  46
  47 #ifdef __UCLIBC_HAS_THREADS__
  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 #undef gettext
  57 #undef gettext_noop
  58 #define gettext(msgid)       (msgid)
  59 #define gettext_noop(String) String
  60
  61 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL)
  62 # define RE_ENABLE_I18N
  63 #endif
  64
  65 #if __GNUC__ >= 3
  66 /* uclibc: lean towards smaller size a bit:
  67  * OFF: # define BE(expr, val) __builtin_expect (expr, val) */
  68 # define BE(expr, val) (expr)
  69 #else
  70 # define BE(expr, val) (expr)
  71 # define inline
  72 #endif
  73
  74 /* Number of single byte character.  */
  75 #define SBC_MAX 256
  76
  77 #define COLL_ELEM_LEN_MAX 8
  78
  79 /* The character which represents newline.  */
  80 #define NEWLINE_CHAR '\n'
  81 #define WIDE_NEWLINE_CHAR L'\n'
  82
  83 #ifdef __GNUC__
  84 # define __attribute(arg) __attribute__ (arg)
  85 #else
  86 # define __attribute(arg)
  87 #endif
  88
  89 /* An integer used to represent a set of bits.  It must be unsigned,
  90    and must be at least as wide as unsigned int.  */
  91 typedef unsigned long int bitset_word_t;
  92 /* All bits set in a bitset_word_t.  */
  93 #define BITSET_WORD_MAX ULONG_MAX
  94 /* Number of bits in a bitset_word_t.  */
  95 #define BITSET_WORD_BITS (sizeof (bitset_word_t) * CHAR_BIT)
  96 /* Number of bitset_word_t in a bit_set.  */
  97 #define BITSET_WORDS (SBC_MAX / BITSET_WORD_BITS)
  98 typedef bitset_word_t bitset_t[BITSET_WORDS];
  99 typedef bitset_word_t *re_bitset_ptr_t;
 100 typedef const bitset_word_t *re_const_bitset_ptr_t;
 101
 102 #define bitset_set(set,i) \
 103   (set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS)
 104 #define bitset_clear(set,i) \
 105   (set[i / BITSET_WORD_BITS] &= ~((bitset_word_t) 1 << i % BITSET_WORD_BITS))
 106 #define bitset_contain(set,i) \
 107   (set[i / BITSET_WORD_BITS] & ((bitset_word_t) 1 << i % BITSET_WORD_BITS))
 108 #define bitset_empty(set) memset (set, '\0', sizeof (bitset_t))
 109 #define bitset_set_all(set) memset (set, '\xff', sizeof (bitset_t))
 110 #define bitset_copy(dest,src) memcpy (dest, src, sizeof (bitset_t))
 111
 112 #define PREV_WORD_CONSTRAINT 0x0001
 113 #define PREV_NOTWORD_CONSTRAINT 0x0002
 114 #define NEXT_WORD_CONSTRAINT 0x0004
 115 #define NEXT_NOTWORD_CONSTRAINT 0x0008
 116 #define PREV_NEWLINE_CONSTRAINT 0x0010
 117 #define NEXT_NEWLINE_CONSTRAINT 0x0020
 118 #define PREV_BEGBUF_CONSTRAINT 0x0040
 119 #define NEXT_ENDBUF_CONSTRAINT 0x0080
 120 #define WORD_DELIM_CONSTRAINT 0x0100
 121 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
 122
 123 typedef enum
 124 {
 125   INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
 126   WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
 127   WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
 128   INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
 129   LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
 130   LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
 131   BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
 132   BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
 133   WORD_DELIM = WORD_DELIM_CONSTRAINT,
 134   NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
 135 } re_context_type;
 136
 137 typedef struct
 138 {
 139   int alloc;
 140   int nelem;
 141   int *elems;
 142 } re_node_set;
 143
 144 typedef enum
 145 {
 146   NON_TYPE = 0,
 147
 148   /* Node type, These are used by token, node, tree.  */
 149   CHARACTER = 1,
 150   END_OF_RE = 2,
 151   SIMPLE_BRACKET = 3,
 152   OP_BACK_REF = 4,
 153   OP_PERIOD = 5,
 154 #ifdef RE_ENABLE_I18N
 155   COMPLEX_BRACKET = 6,
 156   OP_UTF8_PERIOD = 7,
 157 #endif /* RE_ENABLE_I18N */
 158
 159   /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
 160      when the debugger shows values of this enum type.  */
 161 #define EPSILON_BIT 8
 162   OP_OPEN_SUBEXP = EPSILON_BIT | 0,
 163   OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
 164   OP_ALT = EPSILON_BIT | 2,
 165   OP_DUP_ASTERISK = EPSILON_BIT | 3,
 166   ANCHOR = EPSILON_BIT | 4,
 167
 168   /* Tree type, these are used only by tree. */
 169   CONCAT = 16,
 170   SUBEXP = 17,
 171
 172   /* Token type, these are used only by token.  */
 173   OP_DUP_PLUS = 18,
 174   OP_DUP_QUESTION,
 175   OP_OPEN_BRACKET,
 176   OP_CLOSE_BRACKET,
 177   OP_CHARSET_RANGE,
 178   OP_OPEN_DUP_NUM,
 179   OP_CLOSE_DUP_NUM,
 180   OP_NON_MATCH_LIST,
 181   OP_OPEN_COLL_ELEM,
 182   OP_CLOSE_COLL_ELEM,
 183   OP_OPEN_EQUIV_CLASS,
 184   OP_CLOSE_EQUIV_CLASS,
 185   OP_OPEN_CHAR_CLASS,
 186   OP_CLOSE_CHAR_CLASS,
 187   OP_WORD,
 188   OP_NOTWORD,
 189   OP_SPACE,
 190   OP_NOTSPACE,
 191   BACK_SLASH
 192
 193 } re_token_type_t;
 194
 195 #ifdef RE_ENABLE_I18N
 196 typedef struct
 197 {
 198   /* Multibyte characters.  */
 199   wchar_t *mbchars;
 200
 201   /* Collating symbols.  */
 202 # if 0
 203   int32_t *coll_syms;
 204 # endif
 205
 206   /* Equivalence classes. */
 207 # if 0
 208   int32_t *equiv_classes;
 209 # endif
 210
 211   /* Range expressions. */
 212 # if 0
 213   uint32_t *range_starts;
 214   uint32_t *range_ends;
 215 # else
 216   wchar_t *range_starts;
 217   wchar_t *range_ends;
 218 # endif
 219
 220   /* Character classes. */
 221   wctype_t *char_classes;
 222
 223   /* If this character set is the non-matching list.  */
 224   unsigned int non_match : 1;
 225
 226   /* # of multibyte characters.  */
 227   int nmbchars;
 228
 229   /* # of collating symbols.  */
 230   int ncoll_syms;
 231
 232   /* # of equivalence classes. */
 233   int nequiv_classes;
 234
 235   /* # of range expressions. */
 236   int nranges;
 237
 238   /* # of character classes. */
 239   int nchar_classes;
 240 } re_charset_t;
 241 #endif /* RE_ENABLE_I18N */
 242
 243 typedef struct
 244 {
 245   union
 246   {
 247     unsigned char c;            /* for CHARACTER */
 248     re_bitset_ptr_t sbcset;     /* for SIMPLE_BRACKET */
 249 #ifdef RE_ENABLE_I18N
 250     re_charset_t *mbcset;       /* for COMPLEX_BRACKET */
 251 #endif /* RE_ENABLE_I18N */
 252     int idx;                    /* for BACK_REF */
 253     re_context_type ctx_type;   /* for ANCHOR */
 254   } opr;
 255 #if __GNUC__ >= 2
 256   re_token_type_t type : 8;
 257 #else
 258   re_token_type_t type;
 259 #endif
 260   unsigned int constraint : 10; /* context constraint */
 261   unsigned int duplicated : 1;
 262   unsigned int opt_subexp : 1;
 263 #ifdef RE_ENABLE_I18N
 264   unsigned int accept_mb : 1;
 265   /* These 2 bits can be moved into the union if needed (e.g. if running out
 266      of bits; move opr.c to opr.c.c and move the flags to opr.c.flags).  */
 267   unsigned int mb_partial : 1;
 268 #endif
 269   unsigned int word_char : 1;
 270 } re_token_t;
 271
 272 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
 273
 274 struct re_string_t
 275 {
 276   /* Indicate the raw buffer which is the original string passed as an
 277      argument of regexec(), re_search(), etc..  */
 278   const unsigned char *raw_mbs;
 279   /* Store the multibyte string.  In case of "case insensitive mode" like
 280      REG_ICASE, upper cases of the string are stored, otherwise MBS points
 281      the same address that RAW_MBS points.  */
 282   unsigned char *mbs;
 283 #ifdef RE_ENABLE_I18N
 284   /* Store the wide character string which is corresponding to MBS.  */
 285   wint_t *wcs;
 286   int *offsets;
 287   mbstate_t cur_state;
 288 #endif
 289   /* Index in RAW_MBS.  Each character mbs[i] corresponds to
 290      raw_mbs[raw_mbs_idx + i].  */
 291   int raw_mbs_idx;
 292   /* The length of the valid characters in the buffers.  */
 293   int valid_len;
 294   /* The corresponding number of bytes in raw_mbs array.  */
 295   int valid_raw_len;
 296   /* The length of the buffers MBS and WCS.  */
 297   int bufs_len;
 298   /* The index in MBS, which is updated by re_string_fetch_byte.  */
 299   int cur_idx;
 300   /* length of RAW_MBS array.  */
 301   int raw_len;
 302   /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN.  */
 303   int len;
 304   /* End of the buffer may be shorter than its length in the cases such
 305      as re_match_2, re_search_2.  Then, we use STOP for end of the buffer
 306      instead of LEN.  */
 307   int raw_stop;
 308   /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS.  */
 309   int stop;
 310
 311   /* The context of mbs[0].  We store the context independently, since
 312      the context of mbs[0] may be different from raw_mbs[0], which is
 313      the beginning of the input string.  */
 314   unsigned int tip_context;
 315   /* The translation passed as a part of an argument of re_compile_pattern.  */
 316   __RE_TRANSLATE_TYPE trans;
 317   /* Copy of re_dfa_t's word_char.  */
 318   re_const_bitset_ptr_t word_char;
 319   /* 1 if REG_ICASE.  */
 320   unsigned char icase;
 321   unsigned char is_utf8;
 322   unsigned char map_notascii;
 323   unsigned char mbs_allocated;
 324   unsigned char offsets_needed;
 325   unsigned char newline_anchor;
 326   unsigned char word_ops_used;
 327   int mb_cur_max;
 328 };
 329 typedef struct re_string_t re_string_t;
 330
 331 struct re_dfa_t;
 332 typedef struct re_dfa_t re_dfa_t;
 333
 334 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
 335                                                 int new_buf_len)
 336      internal_function;
 337 #ifdef RE_ENABLE_I18N
 338 static void build_wcs_buffer (re_string_t *pstr) internal_function;
 339 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr) internal_function;
 340 #endif /* RE_ENABLE_I18N */
 341 static void build_upper_buffer (re_string_t *pstr) internal_function;
 342 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
 343 static unsigned int re_string_context_at (const re_string_t *input, int idx,
 344                                           int eflags)
 345      internal_function __attribute ((pure));
 346 #define re_string_peek_byte(pstr, offset) \
 347   ((pstr)->mbs[(pstr)->cur_idx + offset])
 348 #define re_string_fetch_byte(pstr) \
 349   ((pstr)->mbs[(pstr)->cur_idx++])
 350 #define re_string_first_byte(pstr, idx) \
 351   ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
 352 #define re_string_is_single_byte_char(pstr, idx) \
 353   ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
 354                                 || (pstr)->wcs[(idx) + 1] != WEOF))
 355 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
 356 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
 357 #define re_string_get_buffer(pstr) ((pstr)->mbs)
 358 #define re_string_length(pstr) ((pstr)->len)
 359 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
 360 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
 361 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
 362
 363 #include <alloca.h>
 364
 365 #if 1
 366 # ifdef HAVE_ALLOCA
 367 /* The OS usually guarantees only one guard page at the bottom of the stack,
 368    and a page size can be as small as 4096 bytes.  So we cannot safely
 369    allocate anything larger than 4096 bytes.  Also care for the possibility
 370    of a few compiler-allocated temporary stack slots.  */
 371 #  define __libc_use_alloca(n) ((n) < 4032)
 372 # else
 373 /* alloca is implemented with malloc, so just use malloc.  */
 374 #  define __libc_use_alloca(n) 0
 375 # endif
 376 #endif
 377
 378 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
 379 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
 380 #define re_free(p) free (p)
 381
 382 struct bin_tree_t
 383 {
 384   struct bin_tree_t *parent;
 385   struct bin_tree_t *left;
 386   struct bin_tree_t *right;
 387   struct bin_tree_t *first;
 388   struct bin_tree_t *next;
 389
 390   re_token_t token;
 391
 392   /* `node_idx' is the index in dfa->nodes, if `type' == 0.
 393      Otherwise `type' indicate the type of this node.  */
 394   int node_idx;
 395 };
 396 typedef struct bin_tree_t bin_tree_t;
 397
 398 #define BIN_TREE_STORAGE_SIZE \
 399   ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
 400
 401 struct bin_tree_storage_t
 402 {
 403   struct bin_tree_storage_t *next;
 404   bin_tree_t data[BIN_TREE_STORAGE_SIZE];
 405 };
 406 typedef struct bin_tree_storage_t bin_tree_storage_t;
 407
 408 #define CONTEXT_WORD 1
 409 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
 410 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
 411 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
 412
 413 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
 414 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
 415 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
 416 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
 417 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
 418
 419 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
 420 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
 421 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
 422 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
 423
 424 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
 425  ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
 426   || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
 427   || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
 428   || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
 429
 430 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
 431  ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
 432   || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
 433   || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
 434   || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
 435
 436 struct re_dfastate_t
 437 {
 438   unsigned int hash;
 439   re_node_set nodes;
 440   re_node_set non_eps_nodes;
 441   re_node_set inveclosure;
 442   re_node_set *entrance_nodes;
 443   struct re_dfastate_t **trtable, **word_trtable;
 444   unsigned int context : 4;
 445   unsigned int halt : 1;
 446   /* If this state can accept `multi byte'.
 447      Note that we refer to multibyte characters, and multi character
 448      collating elements as `multi byte'.  */
 449   unsigned int accept_mb : 1;
 450   /* If this state has backreference node(s).  */
 451   unsigned int has_backref : 1;
 452   unsigned int has_constraint : 1;
 453 };
 454 typedef struct re_dfastate_t re_dfastate_t;
 455
 456 struct re_state_table_entry
 457 {
 458   int num;
 459   int alloc;
 460   re_dfastate_t **array;
 461 };
 462
 463 /* Array type used in re_sub_match_last_t and re_sub_match_top_t.  */
 464
 465 typedef struct
 466 {
 467   int next_idx;
 468   int alloc;
 469   re_dfastate_t **array;
 470 } state_array_t;
 471
 472 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP.  */
 473
 474 typedef struct
 475 {
 476   int node;
 477   int str_idx; /* The position NODE match at.  */
 478   state_array_t path;
 479 } re_sub_match_last_t;
 480
 481 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
 482    And information about the node, whose type is OP_CLOSE_SUBEXP,
 483    corresponding to NODE is stored in LASTS.  */
 484
 485 typedef struct
 486 {
 487   int str_idx;
 488   int node;
 489   state_array_t *path;
 490   int alasts; /* Allocation size of LASTS.  */
 491   int nlasts; /* The number of LASTS.  */
 492   re_sub_match_last_t **lasts;
 493 } re_sub_match_top_t;
 494
 495 struct re_backref_cache_entry
 496 {
 497   int node;
 498   int str_idx;
 499   int subexp_from;
 500   int subexp_to;
 501   char more;
 502   char unused;
 503   unsigned short int eps_reachable_subexps_map;
 504 };
 505
 506 typedef struct
 507 {
 508   /* The string object corresponding to the input string.  */
 509   re_string_t input;
 510   const re_dfa_t *dfa;
 511   /* EFLAGS of the argument of regexec.  */
 512   int eflags;
 513   /* Where the matching ends.  */
 514   int match_last;
 515   int last_node;
 516   /* The state log used by the matcher.  */
 517   re_dfastate_t **state_log;
 518   int state_log_top;
 519   /* Back reference cache.  */
 520   int nbkref_ents;
 521   int abkref_ents;
 522   struct re_backref_cache_entry *bkref_ents;
 523   int max_mb_elem_len;
 524   int nsub_tops;
 525   int asub_tops;
 526   re_sub_match_top_t **sub_tops;
 527 } re_match_context_t;
 528
 529 typedef struct
 530 {
 531   re_dfastate_t **sifted_states;
 532   re_dfastate_t **limited_states;
 533   int last_node;
 534   int last_str_idx;
 535   re_node_set limits;
 536 } re_sift_context_t;
 537
 538 struct re_fail_stack_ent_t
 539 {
 540   int idx;
 541   int node;
 542   regmatch_t *regs;
 543   re_node_set eps_via_nodes;
 544 };
 545
 546 struct re_fail_stack_t
 547 {
 548   int num;
 549   int alloc;
 550   struct re_fail_stack_ent_t *stack;
 551 };
 552
 553 struct re_dfa_t
 554 {
 555   re_token_t *nodes;
 556   size_t nodes_alloc;
 557   size_t nodes_len;
 558   int *nexts;
 559   int *org_indices;
 560   re_node_set *edests;
 561   re_node_set *eclosures;
 562   re_node_set *inveclosures;
 563   struct re_state_table_entry *state_table;
 564   re_dfastate_t *init_state;
 565   re_dfastate_t *init_state_word;
 566   re_dfastate_t *init_state_nl;
 567   re_dfastate_t *init_state_begbuf;
 568   bin_tree_t *str_tree;
 569   bin_tree_storage_t *str_tree_storage;
 570   re_bitset_ptr_t sb_char;
 571   int str_tree_storage_idx;
 572
 573   /* number of subexpressions `re_nsub' is in regex_t.  */
 574   unsigned int state_hash_mask;
 575   int init_node;
 576   int nbackref; /* The number of backreference in this dfa.  */
 577
 578   /* Bitmap expressing which backreference is used.  */
 579   bitset_word_t used_bkref_map;
 580   bitset_word_t completed_bkref_map;
 581
 582   unsigned int has_plural_match : 1;
 583   /* If this dfa has "multibyte node", which is a backreference or
 584      a node which can accept multibyte character or multi character
 585      collating element.  */
 586   unsigned int has_mb_node : 1;
 587   unsigned int is_utf8 : 1;
 588   unsigned int map_notascii : 1;
 589   unsigned int word_ops_used : 1;
 590   int mb_cur_max;
 591   bitset_t word_char;
 592   reg_syntax_t syntax;
 593   int *subexp_map;
 594 #ifdef DEBUG
 595   char* re_str;
 596 #endif
 597   __libc_lock_define (, lock)
 598 };
 599
 600 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
 601 #define re_node_set_remove(set,id) \
 602   (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
 603 #define re_node_set_empty(p) ((p)->nelem = 0)
 604 #define re_node_set_free(set) re_free ((set)->elems)
 605
 606
 607 typedef enum
 608 {
 609   SB_CHAR,
 610   MB_CHAR,
 611   EQUIV_CLASS,
 612   COLL_SYM,
 613   CHAR_CLASS
 614 } bracket_elem_type;
 615
 616 typedef struct
 617 {
 618   bracket_elem_type type;
 619   union
 620   {
 621     unsigned char ch;
 622     unsigned char *name;
 623 #ifdef __UCLIBC_HAS_WCHAR__
 624     wchar_t wch;
 625 #endif
 626   } opr;
 627 } bracket_elem_t;
 628
 629
 630 /* Inline functions for bitset operation.  */
 631 static __inline__ void
 632 bitset_not (bitset_t set)
 633 {
 634   int bitset_i;
 635   for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
 636     set[bitset_i] = ~set[bitset_i];
 637 }
 638
 639 static __inline__ void
 640 bitset_merge (bitset_t dest, const bitset_t src)
 641 {
 642   int bitset_i;
 643   for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
 644     dest[bitset_i] |= src[bitset_i];
 645 }
 646
 647 static __inline__ void
 648 bitset_mask (bitset_t dest, const bitset_t src)
 649 {
 650   int bitset_i;
 651   for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
 652     dest[bitset_i] &= src[bitset_i];
 653 }
 654
 655 #ifdef RE_ENABLE_I18N
 656 /* Inline functions for re_string.  */
 657 static __inline__ int
 658 internal_function __attribute ((pure))
 659 re_string_char_size_at (const re_string_t *pstr, int idx)
 660 {
 661   int byte_idx;
 662   if (pstr->mb_cur_max == 1)
 663     return 1;
 664   for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
 665     if (pstr->wcs[idx + byte_idx] != WEOF)
 666       break;
 667   return byte_idx;
 668 }
 669
 670 static __inline__ wint_t
 671 internal_function __attribute ((pure))
 672 re_string_wchar_at (const re_string_t *pstr, int idx)
 673 {
 674   if (pstr->mb_cur_max == 1)
 675     return (wint_t) pstr->mbs[idx];
 676   return (wint_t) pstr->wcs[idx];
 677 }
 678
 679 static int
 680 internal_function __attribute ((pure))
 681 re_string_elem_size_at (const re_string_t *pstr, int idx)
 682 {
 683 # if 0
 684   const unsigned char *p, *extra;
 685   const int32_t *table, *indirect;
 686   int32_t tmp;
 687 #  include <locale/weight.h>
 688   uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
 689
 690   if (nrules != 0)
 691     {
 692       table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
 693       extra = (const unsigned char *)
 694         _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
 695       indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
 696                                                 _NL_COLLATE_INDIRECTMB);
 697       p = pstr->mbs + idx;
 698       tmp = findidx (&p);
 699       return p - pstr->mbs - idx;
 700     }
 701 # endif
 702   return 1;
 703 }
 704 #endif /* RE_ENABLE_I18N */
 705
 706 #endif /*  _REGEX_INTERNAL_H */