posix/regex_internal.h

   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>.
   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, write to the Free
  18    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  19    02111-1307 USA.  */
  20
  21 #ifndef _REGEX_INTERNAL_H
  22 #define _REGEX_INTERNAL_H 1
  23
  24 #include <assert.h>
  25 #include <ctype.h>
  26 #include <stdio.h>
  27 #include <stdlib.h>
  28 #include <string.h>
  29
  30 #if defined HAVE_LOCALE_H || defined _LIBC
  31 # include <locale.h>
  32 #endif
  33 #if defined HAVE_WCHAR_H || defined _LIBC
  34 # include <wchar.h>
  35 #endif /* HAVE_WCHAR_H || _LIBC */
  36 #if defined HAVE_WCTYPE_H || defined _LIBC
  37 # include <wctype.h>
  38 #endif /* HAVE_WCTYPE_H || _LIBC */
  39
  40 /* In case that the system doesn't have isblank().  */
  41 #if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank
  42 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
  43 #endif
  44
  45 #ifdef _LIBC
  46 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
  47 #  define _RE_DEFINE_LOCALE_FUNCTIONS 1
  48 #   include <locale/localeinfo.h>
  49 #   include <locale/elem-hash.h>
  50 #   include <locale/coll-lookup.h>
  51 # endif
  52 #endif
  53
  54 /* This is for other GNU distributions with internationalized messages.  */
  55 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
  56 # include <libintl.h>
  57 # ifdef _LIBC
  58 #  undef gettext
  59 #  define gettext(msgid) \
  60   INTUSE(__dcgettext) (INTUSE(_libc_intl_domainname), msgid, LC_MESSAGES)
  61 # endif
  62 #else
  63 # define gettext(msgid) (msgid)
  64 #endif
  65
  66 #ifndef gettext_noop
  67 /* This define is so xgettext can find the internationalizable
  68    strings.  */
  69 # define gettext_noop(String) String
  70 #endif
  71
  72 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC
  73 # define RE_ENABLE_I18N
  74 #endif
  75
  76 #if __GNUC__ >= 3
  77 # define BE(expr, val) __builtin_expect (expr, val)
  78 #else
  79 # define BE(expr, val) (expr)
  80 # define inline
  81 #endif
  82
  83 /* Number of bits in a byte.  */
  84 #define BYTE_BITS 8
  85 /* Number of single byte character.  */
  86 #define SBC_MAX 256
  87
  88 #define COLL_ELEM_LEN_MAX 8
  89
  90 /* The character which represents newline.  */
  91 #define NEWLINE_CHAR '\n'
  92 #define WIDE_NEWLINE_CHAR L'\n'
  93
  94 /* Rename to standard API for using out of glibc.  */
  95 #ifndef _LIBC
  96 # define __wctype wctype
  97 # define __iswctype iswctype
  98 # define __btowc btowc
  99 # define __mempcpy mempcpy
 100 # define __wcrtomb wcrtomb
 101 # define attribute_hidden
 102 #endif /* not _LIBC */
 103
 104 #ifdef __GNUC__
 105 # define __attribute(arg) __attribute__ (arg)
 106 #else
 107 # define __attribute(arg)
 108 #endif
 109
 110 extern const char __re_error_msgid[] attribute_hidden;
 111 extern const size_t __re_error_msgid_idx[] attribute_hidden;
 112
 113 /* Number of bits in an unsinged int.  */
 114 #define UINT_BITS (sizeof (unsigned int) * BYTE_BITS)
 115 /* Number of unsigned int in an bit_set.  */
 116 #define BITSET_UINTS ((SBC_MAX + UINT_BITS - 1) / UINT_BITS)
 117 typedef unsigned int bitset[BITSET_UINTS];
 118 typedef unsigned int *re_bitset_ptr_t;
 119 typedef const unsigned int *re_const_bitset_ptr_t;
 120
 121 #define bitset_set(set,i) (set[i / UINT_BITS] |= 1 << i % UINT_BITS)
 122 #define bitset_clear(set,i) (set[i / UINT_BITS] &= ~(1 << i % UINT_BITS))
 123 #define bitset_contain(set,i) (set[i / UINT_BITS] & (1 << i % UINT_BITS))
 124 #define bitset_empty(set) memset (set, 0, sizeof (unsigned int) * BITSET_UINTS)
 125 #define bitset_set_all(set) \
 126   memset (set, 255, sizeof (unsigned int) * BITSET_UINTS)
 127 #define bitset_copy(dest,src) \
 128   memcpy (dest, src, sizeof (unsigned int) * BITSET_UINTS)
 129 static inline void bitset_not (bitset set);
 130 static inline void bitset_merge (bitset dest, const bitset src);
 131 static inline void bitset_not_merge (bitset dest, const bitset src);
 132 static inline void bitset_mask (bitset dest, const bitset src);
 133
 134 #define PREV_WORD_CONSTRAINT 0x0001
 135 #define PREV_NOTWORD_CONSTRAINT 0x0002
 136 #define NEXT_WORD_CONSTRAINT 0x0004
 137 #define NEXT_NOTWORD_CONSTRAINT 0x0008
 138 #define PREV_NEWLINE_CONSTRAINT 0x0010
 139 #define NEXT_NEWLINE_CONSTRAINT 0x0020
 140 #define PREV_BEGBUF_CONSTRAINT 0x0040
 141 #define NEXT_ENDBUF_CONSTRAINT 0x0080
 142 #define DUMMY_CONSTRAINT 0x0100
 143
 144 typedef enum
 145 {
 146   INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
 147   WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
 148   WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
 149   LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
 150   LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
 151   BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
 152   BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
 153   WORD_DELIM = DUMMY_CONSTRAINT
 154 } re_context_type;
 155
 156 typedef struct
 157 {
 158   int alloc;
 159   int nelem;
 160   int *elems;
 161 } re_node_set;
 162
 163 typedef enum
 164 {
 165   NON_TYPE = 0,
 166
 167   /* Node type, These are used by token, node, tree.  */
 168   CHARACTER = 1,
 169   END_OF_RE = 2,
 170   SIMPLE_BRACKET = 3,
 171   OP_BACK_REF = 4,
 172   OP_PERIOD = 5,
 173 #ifdef RE_ENABLE_I18N
 174   COMPLEX_BRACKET = 6,
 175   OP_UTF8_PERIOD = 7,
 176 #endif /* RE_ENABLE_I18N */
 177
 178   /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
 179      when the debugger shows values of this enum type.  */
 180 #define EPSILON_BIT 8
 181   OP_OPEN_SUBEXP = EPSILON_BIT | 0,
 182   OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
 183   OP_ALT = EPSILON_BIT | 2,
 184   OP_DUP_ASTERISK = EPSILON_BIT | 3,
 185   OP_DUP_PLUS = EPSILON_BIT | 4,
 186   OP_DUP_QUESTION = EPSILON_BIT | 5,
 187   ANCHOR = EPSILON_BIT | 6,
 188
 189   /* Tree type, these are used only by tree. */
 190   CONCAT = 16,
 191
 192   /* Token type, these are used only by token.  */
 193   OP_OPEN_BRACKET = 17,
 194   OP_CLOSE_BRACKET,
 195   OP_CHARSET_RANGE,
 196   OP_OPEN_DUP_NUM,
 197   OP_CLOSE_DUP_NUM,
 198   OP_NON_MATCH_LIST,
 199   OP_OPEN_COLL_ELEM,
 200   OP_CLOSE_COLL_ELEM,
 201   OP_OPEN_EQUIV_CLASS,
 202   OP_CLOSE_EQUIV_CLASS,
 203   OP_OPEN_CHAR_CLASS,
 204   OP_CLOSE_CHAR_CLASS,
 205   OP_WORD,
 206   OP_NOTWORD,
 207   OP_SPACE,
 208   OP_NOTSPACE,
 209   BACK_SLASH
 210
 211 } re_token_type_t;
 212
 213 #ifdef RE_ENABLE_I18N
 214 typedef struct
 215 {
 216   /* Multibyte characters.  */
 217   wchar_t *mbchars;
 218
 219   /* Collating symbols.  */
 220 # ifdef _LIBC
 221   int32_t *coll_syms;
 222 # endif
 223
 224   /* Equivalence classes. */
 225 # ifdef _LIBC
 226   int32_t *equiv_classes;
 227 # endif
 228
 229   /* Range expressions. */
 230 # ifdef _LIBC
 231   uint32_t *range_starts;
 232   uint32_t *range_ends;
 233 # else /* not _LIBC */
 234   wchar_t *range_starts;
 235   wchar_t *range_ends;
 236 # endif /* not _LIBC */
 237
 238   /* Character classes. */
 239   wctype_t *char_classes;
 240
 241   /* If this character set is the non-matching list.  */
 242   unsigned int non_match : 1;
 243
 244   /* # of multibyte characters.  */
 245   int nmbchars;
 246
 247   /* # of collating symbols.  */
 248   int ncoll_syms;
 249
 250   /* # of equivalence classes. */
 251   int nequiv_classes;
 252
 253   /* # of range expressions. */
 254   int nranges;
 255
 256   /* # of character classes. */
 257   int nchar_classes;
 258 } re_charset_t;
 259 #endif /* RE_ENABLE_I18N */
 260
 261 typedef struct
 262 {
 263   union
 264   {
 265     unsigned char c;            /* for CHARACTER */
 266     re_bitset_ptr_t sbcset;     /* for SIMPLE_BRACKET */
 267 #ifdef RE_ENABLE_I18N
 268     re_charset_t *mbcset;       /* for COMPLEX_BRACKET */
 269 #endif /* RE_ENABLE_I18N */
 270     int idx;                    /* for BACK_REF */
 271     re_context_type ctx_type;   /* for ANCHOR */
 272   } opr;
 273 #if __GNUC__ >= 2
 274   re_token_type_t type : 8;
 275 #else
 276   re_token_type_t type;
 277 #endif
 278   unsigned int constraint : 10; /* context constraint */
 279   unsigned int duplicated : 1;
 280   unsigned int opt_subexp : 1;
 281 #ifdef RE_ENABLE_I18N
 282   /* These 2 bits can be moved into the union if needed (e.g. if running out
 283      of bits; move opr.c to opr.c.c and move the flags to opr.c.flags).  */
 284   unsigned int mb_partial : 1;
 285 #endif
 286   unsigned int word_char : 1;
 287 } re_token_t;
 288
 289 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
 290 #define ACCEPT_MB_NODE(type) \
 291   ((type) >= OP_PERIOD && (type) <= OP_UTF8_PERIOD)
 292
 293 struct re_string_t
 294 {
 295   /* Indicate the raw buffer which is the original string passed as an
 296      argument of regexec(), re_search(), etc..  */
 297   const unsigned char *raw_mbs;
 298   /* Store the multibyte string.  In case of "case insensitive mode" like
 299      REG_ICASE, upper cases of the string are stored, otherwise MBS points
 300      the same address that RAW_MBS points.  */
 301   unsigned char *mbs;
 302 #ifdef RE_ENABLE_I18N
 303   /* Store the wide character string which is corresponding to MBS.  */
 304   wint_t *wcs;
 305   int *offsets;
 306   mbstate_t cur_state;
 307 #endif
 308   /* Index in RAW_MBS.  Each character mbs[i] corresponds to
 309      raw_mbs[raw_mbs_idx + i].  */
 310   int raw_mbs_idx;
 311   /* The length of the valid characters in the buffers.  */
 312   int valid_len;
 313   /* The corresponding number of bytes in raw_mbs array.  */
 314   int valid_raw_len;
 315   /* The length of the buffers MBS and WCS.  */
 316   int bufs_len;
 317   /* The index in MBS, which is updated by re_string_fetch_byte.  */
 318   int cur_idx;
 319   /* length of RAW_MBS array.  */
 320   int raw_len;
 321   /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN.  */
 322   int len;
 323   /* End of the buffer may be shorter than its length in the cases such
 324      as re_match_2, re_search_2.  Then, we use STOP for end of the buffer
 325      instead of LEN.  */
 326   int raw_stop;
 327   /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS.  */
 328   int stop;
 329
 330   /* The context of mbs[0].  We store the context independently, since
 331      the context of mbs[0] may be different from raw_mbs[0], which is
 332      the beginning of the input string.  */
 333   unsigned int tip_context;
 334   /* The translation passed as a part of an argument of re_compile_pattern.  */
 335   unsigned RE_TRANSLATE_TYPE trans;
 336   /* Copy of re_dfa_t's word_char.  */
 337   re_const_bitset_ptr_t word_char;
 338   /* 1 if REG_ICASE.  */
 339   unsigned char icase;
 340   unsigned char is_utf8;
 341   unsigned char map_notascii;
 342   unsigned char mbs_allocated;
 343   unsigned char offsets_needed;
 344   unsigned char newline_anchor;
 345   unsigned char word_ops_used;
 346   int mb_cur_max;
 347 };
 348 typedef struct re_string_t re_string_t;
 349
 350
 351 struct re_dfa_t;
 352 typedef struct re_dfa_t re_dfa_t;
 353
 354 #ifndef _LIBC
 355 # ifdef __i386__
 356 #  define internal_function   __attribute ((regparm (3), stdcall))
 357 # else
 358 #  define internal_function
 359 # endif
 360 #endif
 361
 362 #ifndef RE_NO_INTERNAL_PROTOTYPES
 363 static reg_errcode_t re_string_allocate (re_string_t *pstr, const char *str,
 364                                          int len, int init_len,
 365                                          RE_TRANSLATE_TYPE trans, int icase,
 366                                          const re_dfa_t *dfa)
 367      internal_function;
 368 static reg_errcode_t re_string_construct (re_string_t *pstr, const char *str,
 369                                           int len, RE_TRANSLATE_TYPE trans,
 370                                           int icase, const re_dfa_t *dfa)
 371      internal_function;
 372 static reg_errcode_t re_string_reconstruct (re_string_t *pstr, int idx,
 373                                             int eflags) internal_function;
 374 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
 375                                                 int new_buf_len)
 376      internal_function;
 377 # ifdef RE_ENABLE_I18N
 378 static void build_wcs_buffer (re_string_t *pstr) internal_function;
 379 static int build_wcs_upper_buffer (re_string_t *pstr) internal_function;
 380 # endif /* RE_ENABLE_I18N */
 381 static void build_upper_buffer (re_string_t *pstr) internal_function;
 382 static void re_string_translate_buffer (re_string_t *pstr) internal_function;
 383 static void re_string_destruct (re_string_t *pstr) internal_function;
 384 # ifdef RE_ENABLE_I18N
 385 static int re_string_elem_size_at (const re_string_t *pstr, int idx)
 386      internal_function;
 387 static inline int re_string_char_size_at (const re_string_t *pstr, int idx)
 388      internal_function;
 389 static inline wint_t re_string_wchar_at (const re_string_t *pstr, int idx)
 390      internal_function;
 391 # endif /* RE_ENABLE_I18N */
 392 static unsigned int re_string_context_at (const re_string_t *input, int idx,
 393                                           int eflags) internal_function;
 394 static unsigned char re_string_peek_byte_case (const re_string_t *pstr,
 395                                                int idx) internal_function;
 396 static unsigned char re_string_fetch_byte_case (re_string_t *pstr)
 397      internal_function;
 398 #endif
 399 #define re_string_peek_byte(pstr, offset) \
 400   ((pstr)->mbs[(pstr)->cur_idx + offset])
 401 #define re_string_fetch_byte(pstr) \
 402   ((pstr)->mbs[(pstr)->cur_idx++])
 403 #define re_string_first_byte(pstr, idx) \
 404   ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
 405 #define re_string_is_single_byte_char(pstr, idx) \
 406   ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
 407                                 || (pstr)->wcs[(idx) + 1] != WEOF))
 408 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
 409 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
 410 #define re_string_get_buffer(pstr) ((pstr)->mbs)
 411 #define re_string_length(pstr) ((pstr)->len)
 412 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
 413 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
 414 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
 415
 416 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
 417 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
 418 #define re_free(p) free (p)
 419
 420 struct bin_tree_t
 421 {
 422   struct bin_tree_t *parent;
 423   struct bin_tree_t *left;
 424   struct bin_tree_t *right;
 425
 426   /* `node_idx' is the index in dfa->nodes, if `type' == 0.
 427      Otherwise `type' indicate the type of this node.  */
 428   re_token_type_t type;
 429   int node_idx;
 430
 431   int first;
 432   int next;
 433   re_node_set eclosure;
 434 };
 435 typedef struct bin_tree_t bin_tree_t;
 436
 437 #define BIN_TREE_STORAGE_SIZE \
 438   ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
 439
 440 struct bin_tree_storage_t
 441 {
 442   struct bin_tree_storage_t *next;
 443   bin_tree_t data[BIN_TREE_STORAGE_SIZE];
 444 };
 445 typedef struct bin_tree_storage_t bin_tree_storage_t;
 446
 447 #define CONTEXT_WORD 1
 448 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
 449 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
 450 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
 451
 452 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
 453 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
 454 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
 455 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
 456 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
 457
 458 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
 459 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
 460 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
 461 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
 462
 463 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
 464  ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
 465   || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
 466   || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
 467   || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
 468
 469 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
 470  ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
 471   || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
 472   || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
 473   || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
 474
 475 struct re_dfastate_t
 476 {
 477   unsigned int hash;
 478   re_node_set nodes;
 479   re_node_set *entrance_nodes;
 480   struct re_dfastate_t **trtable;
 481   unsigned int context : 4;
 482   unsigned int halt : 1;
 483   /* If this state can accept `multi byte'.
 484      Note that we refer to multibyte characters, and multi character
 485      collating elements as `multi byte'.  */
 486   unsigned int accept_mb : 1;
 487   /* If this state has backreference node(s).  */
 488   unsigned int has_backref : 1;
 489   unsigned int has_constraint : 1;
 490   unsigned int word_trtable : 1;
 491 };
 492 typedef struct re_dfastate_t re_dfastate_t;
 493
 494 typedef struct
 495 {
 496   /* start <= node < end  */
 497   int start;
 498   int end;
 499 } re_subexp_t;
 500
 501 struct re_state_table_entry
 502 {
 503   int num;
 504   int alloc;
 505   re_dfastate_t **array;
 506 };
 507
 508 /* Array type used in re_sub_match_last_t and re_sub_match_top_t.  */
 509
 510 typedef struct
 511 {
 512   int next_idx;
 513   int alloc;
 514   re_dfastate_t **array;
 515 } state_array_t;
 516
 517 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP.  */
 518
 519 typedef struct
 520 {
 521   int node;
 522   int str_idx; /* The position NODE match at.  */
 523   state_array_t path;
 524 } re_sub_match_last_t;
 525
 526 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
 527    And information about the node, whose type is OP_CLOSE_SUBEXP,
 528    corresponding to NODE is stored in LASTS.  */
 529
 530 typedef struct
 531 {
 532   int str_idx;
 533   int node;
 534   int next_last_offset;
 535   state_array_t *path;
 536   int alasts; /* Allocation size of LASTS.  */
 537   int nlasts; /* The number of LASTS.  */
 538   re_sub_match_last_t **lasts;
 539 } re_sub_match_top_t;
 540
 541 struct re_backref_cache_entry
 542 {
 543   int node;
 544   int str_idx;
 545   int subexp_from;
 546   int subexp_to;
 547   int flag;
 548 };
 549
 550 typedef struct
 551 {
 552   /* The string object corresponding to the input string.  */
 553   re_string_t input;
 554 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
 555   re_dfa_t *const dfa;
 556 #else
 557   re_dfa_t *dfa;
 558 #endif
 559   /* EFLAGS of the argument of regexec.  */
 560   int eflags;
 561   /* Where the matching ends.  */
 562   int match_last;
 563   int last_node;
 564   /* The state log used by the matcher.  */
 565   re_dfastate_t **state_log;
 566   int state_log_top;
 567   /* Back reference cache.  */
 568   int nbkref_ents;
 569   int abkref_ents;
 570   struct re_backref_cache_entry *bkref_ents;
 571   int max_mb_elem_len;
 572   int nsub_tops;
 573   int asub_tops;
 574   re_sub_match_top_t **sub_tops;
 575 } re_match_context_t;
 576
 577 typedef struct
 578 {
 579   int cur_bkref;
 580   int cls_subexp_idx;
 581
 582   re_dfastate_t **sifted_states;
 583   re_dfastate_t **limited_states;
 584
 585   re_node_set limits;
 586
 587   int last_node;
 588   int last_str_idx;
 589   int check_subexp;
 590 } re_sift_context_t;
 591
 592 struct re_fail_stack_ent_t
 593 {
 594   int idx;
 595   int node;
 596   regmatch_t *regs;
 597   re_node_set eps_via_nodes;
 598 };
 599
 600 struct re_fail_stack_t
 601 {
 602   int num;
 603   int alloc;
 604   struct re_fail_stack_ent_t *stack;
 605 };
 606
 607 struct re_dfa_t
 608 {
 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;
 627
 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
 650 };
 651
 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
 686 \f
 687
 688 typedef enum
 689 {
 690   SB_CHAR,
 691   MB_CHAR,
 692   EQUIV_CLASS,
 693   COLL_SYM,
 694   CHAR_CLASS
 695 } bracket_elem_type;
 696
 697 typedef struct
 698 {
 699   bracket_elem_type type;
 700   union
 701   {
 702     unsigned char ch;
 703     unsigned char *name;
 704     wchar_t wch;
 705   } opr;
 706 } bracket_elem_t;
 707
 708
 709 /* Inline functions for bitset operation.  */
 710 static inline void
 711 bitset_not (bitset set)
 712 {
 713   int bitset_i;
 714   for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i)
 715     set[bitset_i] = ~set[bitset_i];
 716 }
 717
 718 static inline void
 719 bitset_merge (bitset dest, const bitset src)
 720 {
 721   int bitset_i;
 722   for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i)
 723     dest[bitset_i] |= src[bitset_i];
 724 }
 725
 726 static inline void
 727 bitset_not_merge (bitset dest, const bitset src)
 728 {
 729   int i;
 730   for (i = 0; i < BITSET_UINTS; ++i)
 731     dest[i] |= ~src[i];
 732 }
 733
 734 static inline void
 735 bitset_mask (bitset dest, const bitset src)
 736 {
 737   int bitset_i;
 738   for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i)
 739     dest[bitset_i] &= src[bitset_i];
 740 }
 741
 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)
 747 {
 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;
 755 }
 756
 757 static inline wint_t
 758 internal_function
 759 re_string_wchar_at (const re_string_t *pstr, int idx)
 760 {
 761   if (pstr->mb_cur_max == 1)
 762     return (wint_t) pstr->mbs[idx];
 763   return (wint_t) pstr->wcs[idx];
 764 }
 765
 766 static int
 767 internal_function
 768 re_string_elem_size_at (const re_string_t *pstr, int idx)
 769 {
 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);
 776
 777   if (nrules != 0)
 778     {
 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;
 787     }
 788   else
 789 #endif /* _LIBC */
 790     return 1;
 791 }
 792 #endif /* RE_ENABLE_I18N */
 793
 794 #endif /*  _REGEX_INTERNAL_H */