posix/regex_internal.h

   1 /* Extended regular expression matching and search library.
   2    Copyright (C) 2002 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 /* Number of bits in a byte.  */
  25 #define BYTE_BITS 8
  26 /* Number of single byte character.  */
  27 #define SBC_MAX 256
  28
  29 #define COLL_ELEM_LEN_MAX 8
  30
  31 /* The character which represents newline.  */
  32 #define NEWLINE_CHAR '\n'
  33
  34 /* Rename to standard API for using out of glibc.  */
  35 #ifndef _LIBC
  36 # define __wctype wctype
  37 # define __iswctype iswctype
  38 # define __btowc btowc
  39 # define __mempcpy memcpy
  40 # define attribute_hidden
  41 #endif /* not _LIBC */
  42
  43 extern const char __re_error_msgid[] attribute_hidden;
  44 extern const size_t __re_error_msgid_idx[] attribute_hidden;
  45
  46 /* Number of bits in an unsinged int.  */
  47 #define UINT_BITS (sizeof (unsigned int) * BYTE_BITS)
  48 /* Number of unsigned int in an bit_set.  */
  49 #define BITSET_UINTS ((SBC_MAX + UINT_BITS - 1) / UINT_BITS)
  50 typedef unsigned int bitset[BITSET_UINTS];
  51 typedef unsigned int *re_bitset_ptr_t;
  52
  53 #define bitset_set(set,i) (set[i / UINT_BITS] |= 1 << i % UINT_BITS)
  54 #define bitset_clear(set,i) (set[i / UINT_BITS] &= ~(1 << i % UINT_BITS))
  55 #define bitset_contain(set,i) (set[i / UINT_BITS] & (1 << i % UINT_BITS))
  56 #define bitset_empty(set) memset (set, 0, sizeof (unsigned int) * BITSET_UINTS)
  57 #define bitset_set_all(set) \
  58   memset (set, 255, sizeof (unsigned int) * BITSET_UINTS)
  59 #define bitset_copy(dest,src) \
  60   memcpy (dest, src, sizeof (unsigned int) * BITSET_UINTS)
  61 static inline void bitset_not (bitset set);
  62 static inline void bitset_merge (bitset dest, const bitset src);
  63 static inline void bitset_not_merge (bitset dest, const bitset src);
  64
  65 #define PREV_WORD_CONSTRAINT 0x0001
  66 #define PREV_NOTWORD_CONSTRAINT 0x0002
  67 #define NEXT_WORD_CONSTRAINT 0x0004
  68 #define NEXT_NOTWORD_CONSTRAINT 0x0008
  69 #define PREV_NEWLINE_CONSTRAINT 0x0010
  70 #define NEXT_NEWLINE_CONSTRAINT 0x0020
  71 #define PREV_BEGBUF_CONSTRAINT 0x0040
  72 #define NEXT_ENDBUF_CONSTRAINT 0x0080
  73 #define DUMMY_CONSTRAINT 0x0100
  74
  75 typedef enum
  76 {
  77   INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
  78   WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
  79   WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
  80   LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
  81   LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
  82   BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
  83   BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
  84   WORD_DELIM = DUMMY_CONSTRAINT
  85 } re_context_type;
  86
  87 typedef struct
  88 {
  89   int alloc;
  90   int nelem;
  91   int *elems;
  92 } re_node_set;
  93
  94 typedef enum
  95 {
  96   NON_TYPE = 0,
  97
  98   /* Token type, these are used only by token.  */
  99   OP_OPEN_BRACKET,
 100   OP_CLOSE_BRACKET,
 101   OP_CHARSET_RANGE,
 102   OP_OPEN_DUP_NUM,
 103   OP_CLOSE_DUP_NUM,
 104   OP_NON_MATCH_LIST,
 105   OP_OPEN_COLL_ELEM,
 106   OP_CLOSE_COLL_ELEM,
 107   OP_OPEN_EQUIV_CLASS,
 108   OP_CLOSE_EQUIV_CLASS,
 109   OP_OPEN_CHAR_CLASS,
 110   OP_CLOSE_CHAR_CLASS,
 111   OP_WORD,
 112   OP_NOTWORD,
 113   BACK_SLASH,
 114
 115   /* Tree type, these are used only by tree. */
 116   CONCAT,
 117   ALT,
 118   SUBEXP,
 119   SIMPLE_BRACKET,
 120 #ifdef RE_ENABLE_I18N
 121   COMPLEX_BRACKET,
 122 #endif /* RE_ENABLE_I18N */
 123
 124   /* Node type, These are used by token, node, tree.  */
 125   OP_OPEN_SUBEXP,
 126   OP_CLOSE_SUBEXP,
 127   OP_PERIOD,
 128   CHARACTER,
 129   END_OF_RE,
 130   OP_ALT,
 131   OP_DUP_ASTERISK,
 132   OP_DUP_PLUS,
 133   OP_DUP_QUESTION,
 134   OP_BACK_REF,
 135   ANCHOR,
 136   OP_CONTEXT_NODE,
 137
 138   /* Dummy marker.  */
 139   END_OF_RE_TOKEN_T
 140 } re_token_type_t;
 141
 142 #ifdef RE_ENABLE_I18N
 143 typedef struct
 144 {
 145   /* Multibyte characters.  */
 146   wchar_t *mbchars;
 147
 148   /* Collating symbols.  */
 149 # ifdef _LIBC
 150   int32_t *coll_syms;
 151 # endif
 152
 153   /* Equivalence classes. */
 154 # ifdef _LIBC
 155   int32_t *equiv_classes;
 156 # endif
 157
 158   /* Range expressions. */
 159 # ifdef _LIBC
 160   uint32_t *range_starts;
 161   uint32_t *range_ends;
 162 # else /* not _LIBC */
 163   wchar_t *range_starts;
 164   wchar_t *range_ends;
 165 # endif /* not _LIBC */
 166
 167   /* Character classes. */
 168   wctype_t *char_classes;
 169
 170   /* If this character set is the non-matching list.  */
 171   unsigned int non_match : 1;
 172
 173   /* # of multibyte characters.  */
 174   int nmbchars;
 175
 176   /* # of collating symbols.  */
 177   int ncoll_syms;
 178
 179   /* # of equivalence classes. */
 180   int nequiv_classes;
 181
 182   /* # of range expressions. */
 183   int nranges;
 184
 185   /* # of character classes. */
 186   int nchar_classes;
 187 } re_charset_t;
 188 #endif /* RE_ENABLE_I18N */
 189
 190 typedef struct
 191 {
 192   union
 193   {
 194     unsigned char c;            /* for CHARACTER */
 195     re_bitset_ptr_t sbcset;     /* for SIMPLE_BRACKET */
 196 #ifdef RE_ENABLE_I18N
 197     re_charset_t *mbcset;       /* for COMPLEX_BRACKET */
 198 #endif /* RE_ENABLE_I18N */
 199     int idx;                    /* for BACK_REF */
 200     re_context_type ctx_type;   /* for ANCHOR */
 201     struct
 202     {
 203       int entity;               /* for OP_CONTEXT_NODE, index of the entity */
 204       re_node_set *bkref_eclosure;
 205     } *ctx_info;
 206   } opr;
 207 #if __GNUC__ >= 2
 208   re_token_type_t type : 8;
 209 #else
 210   re_token_type_t type;
 211 #endif
 212   unsigned int constraint : 10; /* context constraint */
 213   unsigned int duplicated : 1;
 214 #ifdef RE_ENABLE_I18N
 215   unsigned int mb_partial : 1;
 216 #endif
 217 } re_token_t;
 218
 219 #define IS_EPSILON_NODE(type) \
 220   ((type) == OP_ALT || (type) == OP_DUP_ASTERISK || (type) == OP_DUP_PLUS \
 221    || (type) == OP_DUP_QUESTION || (type) == ANCHOR \
 222    || (type) == OP_OPEN_SUBEXP || (type) == OP_CLOSE_SUBEXP)
 223
 224 #define ACCEPT_MB_NODE(type) \
 225   ((type) == COMPLEX_BRACKET || (type) == OP_PERIOD)
 226
 227 struct re_string_t
 228 {
 229   /* Indicate the raw buffer which is the original string passed as an
 230      argument of regexec(), re_search(), etc..  */
 231   const unsigned char *raw_mbs;
 232   /* Store the multibyte string.  In case of "case insensitive mode" like
 233      REG_ICASE, upper cases of the string are stored, otherwise MBS points
 234      the same address that RAW_MBS points.  */
 235   unsigned char *mbs;
 236   /* Store the case sensitive multibyte string.  In case of
 237      "case insensitive mode", the original string are stored,
 238      otherwise MBS_CASE points the same address that MBS points.  */
 239   unsigned char *mbs_case;
 240 #ifdef RE_ENABLE_I18N
 241   /* Store the wide character string which is corresponding to MBS.  */
 242   wint_t *wcs;
 243   mbstate_t cur_state;
 244 #endif
 245   /* Index in RAW_MBS.  Each character mbs[i] corresponds to
 246      raw_mbs[raw_mbs_idx + i].  */
 247   int raw_mbs_idx;
 248   /* The length of the valid characters in the buffers.  */
 249   int valid_len;
 250   /* The length of the buffers MBS, MBS_CASE, and WCS.  */
 251   int bufs_len;
 252   /* The index in MBS, which is updated by re_string_fetch_byte.  */
 253   int cur_idx;
 254   /* This is length_of_RAW_MBS - RAW_MBS_IDX.  */
 255   int len;
 256   /* End of the buffer may be shorter than its length in the cases such
 257      as re_match_2, re_search_2.  Then, we use STOP for end of the buffer
 258      instead of LEN.  */
 259   int stop;
 260
 261   /* The context of mbs[0].  We store the context independently, since
 262      the context of mbs[0] may be different from raw_mbs[0], which is
 263      the beginning of the input string.  */
 264   unsigned int tip_context;
 265   /* The translation passed as a part of an argument of re_compile_pattern.  */
 266   RE_TRANSLATE_TYPE trans;
 267   /* 1 if REG_ICASE.  */
 268   unsigned int icase : 1;
 269 };
 270 typedef struct re_string_t re_string_t;
 271 /* In case of REG_ICASE, we allocate the buffer dynamically for mbs.  */
 272 #define MBS_ALLOCATED(pstr) (pstr->icase)
 273 /* In case that we need translation, we allocate the buffer dynamically
 274    for mbs_case.  Note that mbs == mbs_case if not REG_ICASE.  */
 275 #define MBS_CASE_ALLOCATED(pstr) (pstr->trans != NULL)
 276
 277
 278 static reg_errcode_t re_string_allocate (re_string_t *pstr, const char *str,
 279                                          int len, int init_len,
 280                                          RE_TRANSLATE_TYPE trans, int icase);
 281 static reg_errcode_t re_string_construct (re_string_t *pstr, const char *str,
 282                                           int len, RE_TRANSLATE_TYPE trans,
 283                                           int icase);
 284 static reg_errcode_t re_string_reconstruct (re_string_t *pstr, int idx,
 285                                             int eflags, int newline);
 286 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
 287                                                 int new_buf_len);
 288 #ifdef RE_ENABLE_I18N
 289 static void build_wcs_buffer (re_string_t *pstr);
 290 static void build_wcs_upper_buffer (re_string_t *pstr);
 291 #endif /* RE_ENABLE_I18N */
 292 static void build_upper_buffer (re_string_t *pstr);
 293 static void re_string_translate_buffer (re_string_t *pstr);
 294 static void re_string_destruct (re_string_t *pstr);
 295 #ifdef RE_ENABLE_I18N
 296 static int re_string_elem_size_at (const re_string_t *pstr, int idx);
 297 static inline int re_string_char_size_at (const re_string_t *pstr, int idx);
 298 static inline wint_t re_string_wchar_at (const re_string_t *pstr, int idx);
 299 #endif /* RE_ENABLE_I18N */
 300 static unsigned int re_string_context_at (const re_string_t *input, int idx,
 301                                           int eflags, int newline_anchor);
 302 #define re_string_peek_byte(pstr, offset) \
 303   ((pstr)->mbs[(pstr)->cur_idx + offset])
 304 #define re_string_peek_byte_case(pstr, offset) \
 305   ((pstr)->mbs_case[(pstr)->cur_idx + offset])
 306 #define re_string_fetch_byte(pstr) \
 307   ((pstr)->mbs[(pstr)->cur_idx++])
 308 #define re_string_fetch_byte_case(pstr) \
 309   ((pstr)->mbs_case[(pstr)->cur_idx++])
 310 #define re_string_first_byte(pstr, idx) \
 311   ((idx) == (pstr)->len || (pstr)->wcs[idx] != WEOF)
 312 #define re_string_is_single_byte_char(pstr, idx) \
 313   ((pstr)->wcs[idx] != WEOF && ((pstr)->len == (idx) \
 314                                 || (pstr)->wcs[(idx) + 1] != WEOF))
 315 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
 316 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
 317 #define re_string_get_buffer(pstr) ((pstr)->mbs)
 318 #define re_string_length(pstr) ((pstr)->len)
 319 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
 320 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
 321 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
 322
 323 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
 324 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
 325 #define re_free(p) free (p)
 326
 327 struct bin_tree_t
 328 {
 329   struct bin_tree_t *parent;
 330   struct bin_tree_t *left;
 331   struct bin_tree_t *right;
 332
 333   /* `node_idx' is the index in dfa->nodes, if `type' == 0.
 334      Otherwise `type' indicate the type of this node.  */
 335   re_token_type_t type;
 336   int node_idx;
 337
 338   int first;
 339   int next;
 340   re_node_set eclosure;
 341 };
 342 typedef struct bin_tree_t bin_tree_t;
 343
 344
 345 #define CONTEXT_WORD 1
 346 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
 347 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
 348 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
 349
 350 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
 351 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
 352 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
 353 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
 354 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
 355
 356 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
 357 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
 358
 359 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
 360  ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
 361   || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
 362   || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
 363   || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
 364
 365 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
 366  ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
 367   || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
 368   || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
 369   || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
 370
 371 struct re_dfastate_t
 372 {
 373   unsigned int hash;
 374   re_node_set nodes;
 375   re_node_set *entrance_nodes;
 376   struct re_dfastate_t **trtable;
 377   struct re_dfastate_t **trtable_search;
 378   /* If this state is a special state.
 379      A state is a special state if the state is the halt state, or
 380      a anchor.  */
 381   unsigned int context : 2;
 382   unsigned int halt : 1;
 383   /* If this state can accept `multi byte'.
 384      Note that we refer to multibyte characters, and multi character
 385      collating elements as `multi byte'.  */
 386   unsigned int accept_mb : 1;
 387   /* If this state has backreference node(s).  */
 388   unsigned int has_backref : 1;
 389   unsigned int has_constraint : 1;
 390 };
 391 typedef struct re_dfastate_t re_dfastate_t;
 392
 393 typedef struct
 394 {
 395   /* start <= node < end  */
 396   int start;
 397   int end;
 398 } re_subexp_t;
 399
 400 struct re_state_table_entry
 401 {
 402   int num;
 403   int alloc;
 404   re_dfastate_t **array;
 405 };
 406
 407 struct re_backref_cache_entry
 408 {
 409   int node;
 410   int from;
 411   int to;
 412   int flag;
 413 };
 414
 415 typedef struct
 416 {
 417   /* EFLAGS of the argument of regexec.  */
 418   int eflags;
 419   /* Where the matching ends.  */
 420   int match_last;
 421   /* The string object corresponding to the input string.  */
 422   re_string_t *input;
 423   /* The state log used by the matcher.  */
 424   re_dfastate_t **state_log;
 425   int state_log_top;
 426   /* Back reference cache.  */
 427   int nbkref_ents;
 428   int abkref_ents;
 429   struct re_backref_cache_entry *bkref_ents;
 430   int max_bkref_len;
 431 } re_match_context_t;
 432
 433 struct re_dfa_t
 434 {
 435   re_bitset_ptr_t word_char;
 436
 437   /* number of subexpressions `re_nsub' is in regex_t.  */
 438   int subexps_alloc;
 439   re_subexp_t *subexps;
 440
 441   re_token_t *nodes;
 442   int nodes_alloc;
 443   int nodes_len;
 444   bin_tree_t *str_tree;
 445   int *firsts;
 446   int *nexts;
 447   re_node_set *edests;
 448   re_node_set *eclosures;
 449   re_node_set *inveclosures;
 450   struct re_state_table_entry *state_table;
 451   unsigned int state_hash_mask;
 452   re_dfastate_t *init_state;
 453   re_dfastate_t *init_state_word;
 454   re_dfastate_t *init_state_nl;
 455   re_dfastate_t *init_state_begbuf;
 456   int states_alloc;
 457   int init_node;
 458   int nbackref; /* The number of backreference in this dfa.  */
 459   /* If this dfa has "multibyte node", which is a backreference or
 460      a node which can accept multibyte character or multi character
 461      collating element.  */
 462   unsigned int has_mb_node : 1;
 463 };
 464 typedef struct re_dfa_t re_dfa_t;
 465
 466 static reg_errcode_t re_node_set_alloc (re_node_set *set, int size);
 467 static reg_errcode_t re_node_set_init_1 (re_node_set *set, int elem);
 468 static reg_errcode_t re_node_set_init_2 (re_node_set *set, int elem1,
 469                                          int elem2);
 470 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
 471 static reg_errcode_t re_node_set_init_copy (re_node_set *dest,
 472                                             const re_node_set *src);
 473 static reg_errcode_t re_node_set_intersect (re_node_set *dest,
 474                                             const re_node_set *src1,
 475                                             const re_node_set *src2);
 476 static reg_errcode_t re_node_set_add_intersect (re_node_set *dest,
 477                                                 const re_node_set *src1,
 478                                                 const re_node_set *src2);
 479 static reg_errcode_t re_node_set_init_union (re_node_set *dest,
 480                                              const re_node_set *src1,
 481                                              const re_node_set *src2);
 482 static reg_errcode_t re_node_set_merge (re_node_set *dest,
 483                                         const re_node_set *src);
 484 static int re_node_set_insert (re_node_set *set, int elem);
 485 static int re_node_set_compare (const re_node_set *set1,
 486                                 const re_node_set *set2);
 487 static int re_node_set_contains (const re_node_set *set, int elem);
 488 static void re_node_set_remove_at (re_node_set *set, int idx);
 489 #define re_node_set_empty(p) ((p)->nelem = 0)
 490 #define re_node_set_free(set) re_free ((set)->elems)
 491 static int re_dfa_add_node (re_dfa_t *dfa, re_token_t token, int mode);
 492 static re_dfastate_t *re_acquire_state (reg_errcode_t *err, re_dfa_t *dfa,
 493                                         const re_node_set *nodes);
 494 static re_dfastate_t *re_acquire_state_context (reg_errcode_t *err,
 495                                                 re_dfa_t *dfa,
 496                                                 const re_node_set *nodes,
 497                                                 unsigned int context);
 498 \f
 499
 500 typedef enum
 501 {
 502   SB_CHAR,
 503   MB_CHAR,
 504   EQUIV_CLASS,
 505   COLL_SYM,
 506   CHAR_CLASS
 507 } bracket_elem_type;
 508
 509 typedef struct
 510 {
 511   bracket_elem_type type;
 512   union
 513   {
 514     unsigned char ch;
 515     unsigned char *name;
 516     wchar_t wch;
 517   } opr;
 518 } bracket_elem_t;
 519
 520
 521 /* Inline functions for bitset operation.  */
 522 static inline void
 523 bitset_not (set)
 524      bitset set;
 525 {
 526   int bitset_i;
 527   for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i)
 528     set[bitset_i] = ~set[bitset_i];
 529 }
 530
 531 static inline void
 532 bitset_merge (dest, src)
 533      bitset dest;
 534      const bitset src;
 535 {
 536   int bitset_i;
 537   for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i)
 538     dest[bitset_i] |= src[bitset_i];
 539 }
 540
 541 static inline void
 542 bitset_not_merge (dest, src)
 543      bitset dest;
 544      const bitset src;
 545 {
 546   int i;
 547   for (i = 0; i < BITSET_UINTS; ++i)
 548     dest[i] |= ~src[i];
 549 }
 550
 551 #ifdef RE_ENABLE_I18N
 552 /* Inline functions for re_string.  */
 553 static inline int
 554 re_string_char_size_at (pstr, idx)
 555      const re_string_t *pstr;
 556      int idx;
 557 {
 558   int byte_idx;
 559   if (MB_CUR_MAX == 1)
 560     return 1;
 561   for (byte_idx = 1; idx + byte_idx < pstr->len; ++byte_idx)
 562     if (pstr->wcs[idx + byte_idx] != WEOF)
 563       break;
 564   return byte_idx;
 565 }
 566
 567 static inline wint_t
 568 re_string_wchar_at (pstr, idx)
 569      const re_string_t *pstr;
 570      int idx;
 571 {
 572   if (MB_CUR_MAX == 1)
 573     return (wint_t) pstr->mbs[idx];
 574   return (wint_t) pstr->wcs[idx];
 575 }
 576
 577 static int
 578 re_string_elem_size_at (pstr, idx)
 579      const re_string_t *pstr;
 580      int idx;
 581 {
 582 #ifdef _LIBC
 583   const unsigned char *p, *extra;
 584   const int32_t *table, *indirect;
 585   int32_t tmp;
 586 # include <locale/weight.h>
 587   uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
 588
 589   if (nrules != 0)
 590     {
 591       table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
 592       extra = (const unsigned char *)
 593         _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
 594       indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
 595                                                 _NL_COLLATE_INDIRECTMB);
 596       p = pstr->mbs + idx;
 597       tmp = findidx (&p);
 598       return p - pstr->mbs - idx;
 599     }
 600   else
 601 #endif /* _LIBC */
 602     return 1;
 603 }
 604 #endif /* RE_ENABLE_I18N */
 605
 606 #endif /*  _REGEX_INTERNAL_H */