1/* Extended regular expression matching and search library. 2 Copyright (C) 2002, 2003 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#ifdef HAVE_CONFIG_H 25#include "config.h" 26#endif 27 28#include <assert.h> 29#include <ctype.h> 30#if 0 31/* Don't include this here. On some systems it sets RE_DUP_MAX to a 32 * lower value than GNU regex allows. Instead, include it in 33 * regex.c, before include of <regex.h>, which correctly 34 * #undefs RE_DUP_MAX and sets it to the right value. 35 */ 36#include <limits.h> 37#endif 38#include <stdio.h> 39#include <stdlib.h> 40#include <string.h> 41 42#if defined HAVE_LOCALE_H || defined _LIBC 43# include <locale.h> 44#endif 45#if defined HAVE_WCHAR_H || defined _LIBC 46# include <wchar.h> 47#endif /* HAVE_WCHAR_H || _LIBC */ 48#if defined HAVE_WCTYPE_H || defined _LIBC 49# include <wctype.h> 50#endif /* HAVE_WCTYPE_H || _LIBC */ 51 52/* In case that the system doesn't have isblank(). */ 53#if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank 54# define isblank(ch) ((ch) == ' ' || (ch) == '\t') 55#endif 56 57#ifdef _LIBC 58# ifndef _RE_DEFINE_LOCALE_FUNCTIONS 59# define _RE_DEFINE_LOCALE_FUNCTIONS 1 60# include <locale/localeinfo.h> 61# include <locale/elem-hash.h> 62# include <locale/coll-lookup.h> 63# endif 64#endif 65 66/* This is for other GNU distributions with internationalized messages. */ 67#if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC 68# include <libintl.h> 69# ifdef _LIBC 70# undef gettext 71# define gettext(msgid) \ 72 INTUSE(__dcgettext) (INTUSE(_libc_intl_domainname), msgid, LC_MESSAGES) 73# endif 74#else 75# define gettext(msgid) (msgid) 76#endif 77 78#ifndef gettext_noop 79/* This define is so xgettext can find the internationalizable 80 strings. */ 81# define gettext_noop(String) String 82#endif 83 84#if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC 85# define RE_ENABLE_I18N 86#endif 87 88#if __GNUC__ >= 3 89# define BE(expr, val) __builtin_expect (expr, val) 90#else 91# define BE(expr, val) (expr) 92# define inline 93#endif 94 95/* Number of bits in a byte. */ 96#define BYTE_BITS 8 97/* Number of single byte character. */ 98#define SBC_MAX 256 99 100#define COLL_ELEM_LEN_MAX 8 101 102/* The character which represents newline. */ 103#define NEWLINE_CHAR '\n' 104#define WIDE_NEWLINE_CHAR L'\n' 105 106/* Rename to standard API for using out of glibc. */ 107#ifndef _LIBC 108# define __wctype wctype 109# define __iswctype iswctype 110# define __btowc btowc 111# define __wcrtomb wcrtomb 112# define attribute_hidden 113# define __thread 114#endif /* not _LIBC */ 115 116#if _LIBC || __GNUC__ >= 3 117# define BE(expr, val) __builtin_expect (expr, val) 118#else 119# define BE(expr, val) (expr) 120# define inline 121#endif 122 123#ifdef RE_ENABLE_I18N 124__thread int re_mb_cur_max = 1; 125#endif 126 127extern const char __re_error_msgid[] attribute_hidden; 128extern const size_t __re_error_msgid_idx[] attribute_hidden; 129 130/* Number of bits in an unsinged int. */ 131#define UINT_BITS (sizeof (unsigned int) * BYTE_BITS) 132/* Number of unsigned int in an bit_set. */ 133#define BITSET_UINTS ((SBC_MAX + UINT_BITS - 1) / UINT_BITS) 134typedef unsigned int bitset[BITSET_UINTS]; 135typedef unsigned int *re_bitset_ptr_t; 136 137#define bitset_set(set,i) (set[i / UINT_BITS] |= 1UL << i % UINT_BITS) 138#define bitset_clear(set,i) (set[i / UINT_BITS] &= ~(1UL << i % UINT_BITS)) 139#define bitset_contain(set,i) (set[i / UINT_BITS] & (1UL << i % UINT_BITS)) 140#define bitset_empty(set) memset (set, 0, sizeof (unsigned int) * BITSET_UINTS) 141#define bitset_set_all(set) \ 142 memset (set, 255, sizeof (unsigned int) * BITSET_UINTS) 143#define bitset_copy(dest,src) \ 144 memcpy (dest, src, sizeof (unsigned int) * BITSET_UINTS) 145static inline void bitset_not _RE_ARGS((bitset set)); 146static inline void bitset_merge _RE_ARGS((bitset dest, const bitset src)); 147static inline void bitset_not_merge _RE_ARGS((bitset dest, const bitset src)); 148 149#define PREV_WORD_CONSTRAINT 0x0001 150#define PREV_NOTWORD_CONSTRAINT 0x0002 151#define NEXT_WORD_CONSTRAINT 0x0004 152#define NEXT_NOTWORD_CONSTRAINT 0x0008 153#define PREV_NEWLINE_CONSTRAINT 0x0010 154#define NEXT_NEWLINE_CONSTRAINT 0x0020 155#define PREV_BEGBUF_CONSTRAINT 0x0040 156#define NEXT_ENDBUF_CONSTRAINT 0x0080 157#define DUMMY_CONSTRAINT 0x0100 158 159typedef enum 160{ 161 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, 162 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, 163 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT, 164 LINE_FIRST = PREV_NEWLINE_CONSTRAINT, 165 LINE_LAST = NEXT_NEWLINE_CONSTRAINT, 166 BUF_FIRST = PREV_BEGBUF_CONSTRAINT, 167 BUF_LAST = NEXT_ENDBUF_CONSTRAINT, 168 WORD_DELIM = DUMMY_CONSTRAINT 169} re_context_type; 170 171typedef struct 172{ 173 int alloc; 174 int nelem; 175 int *elems; 176} re_node_set; 177 178typedef enum 179{ 180 NON_TYPE = 0, 181 182 /* Token type, these are used only by token. */ 183 OP_OPEN_BRACKET, 184 OP_CLOSE_BRACKET, 185 OP_CHARSET_RANGE, 186 OP_OPEN_DUP_NUM, 187 OP_CLOSE_DUP_NUM, 188 OP_NON_MATCH_LIST, 189 OP_OPEN_COLL_ELEM, 190 OP_CLOSE_COLL_ELEM, 191 OP_OPEN_EQUIV_CLASS, 192 OP_CLOSE_EQUIV_CLASS, 193 OP_OPEN_CHAR_CLASS, 194 OP_CLOSE_CHAR_CLASS, 195 OP_WORD, 196 OP_NOTWORD, 197 BACK_SLASH, 198 199 /* Tree type, these are used only by tree. */ 200 CONCAT, 201 ALT, 202 SUBEXP, 203 SIMPLE_BRACKET, 204#ifdef RE_ENABLE_I18N 205 COMPLEX_BRACKET, 206#endif /* RE_ENABLE_I18N */ 207 208 /* Node type, These are used by token, node, tree. */ 209 OP_OPEN_SUBEXP, 210 OP_CLOSE_SUBEXP, 211 OP_PERIOD, 212 CHARACTER, 213 END_OF_RE, 214 OP_ALT, 215 OP_DUP_ASTERISK, 216 OP_DUP_PLUS, 217 OP_DUP_QUESTION, 218 OP_BACK_REF, 219 ANCHOR, 220 221 /* Dummy marker. */ 222 END_OF_RE_TOKEN_T 223} re_token_type_t; 224 225#ifdef RE_ENABLE_I18N 226typedef struct 227{ 228 /* Multibyte characters. */ 229 wchar_t *mbchars; 230 231 /* Collating symbols. */ 232# ifdef _LIBC 233 int32_t *coll_syms; 234# endif 235 236 /* Equivalence classes. */ 237# ifdef _LIBC 238 int32_t *equiv_classes; 239# endif 240 241 /* Range expressions. */ 242# ifdef _LIBC 243 uint32_t *range_starts; 244 uint32_t *range_ends; 245# else /* not _LIBC */ 246 wchar_t *range_starts; 247 wchar_t *range_ends; 248# endif /* not _LIBC */ 249 250 /* Character classes. */ 251 wctype_t *char_classes; 252 253 /* If this character set is the non-matching list. */ 254 unsigned int non_match : 1; 255 256 /* # of multibyte characters. */ 257 int nmbchars; 258 259 /* # of collating symbols. */ 260 int ncoll_syms; 261 262 /* # of equivalence classes. */ 263 int nequiv_classes; 264 265 /* # of range expressions. */ 266 int nranges; 267 268 /* # of character classes. */ 269 int nchar_classes; 270} re_charset_t; 271#endif /* RE_ENABLE_I18N */ 272 273typedef struct 274{ 275 union 276 { 277 unsigned char c; /* for CHARACTER */ 278 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */ 279#ifdef RE_ENABLE_I18N 280 re_charset_t *mbcset; /* for COMPLEX_BRACKET */ 281#endif /* RE_ENABLE_I18N */ 282 int idx; /* for BACK_REF */ 283 re_context_type ctx_type; /* for ANCHOR */ 284 } opr; 285#if __GNUC__ >= 2 286 re_token_type_t type : 8; 287#else 288 re_token_type_t type; 289#endif 290 unsigned int constraint : 10; /* context constraint */ 291 unsigned int duplicated : 1; 292#ifdef RE_ENABLE_I18N 293 unsigned int mb_partial : 1; 294#endif 295} re_token_t; 296 297#define IS_EPSILON_NODE(type) \ 298 ((type) == OP_ALT || (type) == OP_DUP_ASTERISK || (type) == OP_DUP_PLUS \ 299 || (type) == OP_DUP_QUESTION || (type) == ANCHOR \ 300 || (type) == OP_OPEN_SUBEXP || (type) == OP_CLOSE_SUBEXP) 301 302#define ACCEPT_MB_NODE(type) \ 303 ((type) == COMPLEX_BRACKET || (type) == OP_PERIOD) 304 305struct re_string_t 306{ 307 /* Indicate the raw buffer which is the original string passed as an 308 argument of regexec(), re_search(), etc.. */ 309 const unsigned char *raw_mbs; 310 /* Store the multibyte string. In case of "case insensitive mode" like 311 REG_ICASE, upper cases of the string are stored, otherwise MBS points 312 the same address that RAW_MBS points. */ 313 unsigned char *mbs; 314 /* Store the case sensitive multibyte string. In case of 315 "case insensitive mode", the original string are stored, 316 otherwise MBS_CASE points the same address that MBS points. */ 317 unsigned char *mbs_case; 318#ifdef RE_ENABLE_I18N 319 /* Store the wide character string which is corresponding to MBS. */ 320 wint_t *wcs; 321 mbstate_t cur_state; 322#endif 323 /* Index in RAW_MBS. Each character mbs[i] corresponds to 324 raw_mbs[raw_mbs_idx + i]. */ 325 int raw_mbs_idx; 326 /* The length of the valid characters in the buffers. */ 327 int valid_len; 328 /* The length of the buffers MBS, MBS_CASE, and WCS. */ 329 int bufs_len; 330 /* The index in MBS, which is updated by re_string_fetch_byte. */ 331 int cur_idx; 332 /* This is length_of_RAW_MBS - RAW_MBS_IDX. */ 333 int len; 334 /* End of the buffer may be shorter than its length in the cases such 335 as re_match_2, re_search_2. Then, we use STOP for end of the buffer 336 instead of LEN. */ 337 int stop; 338 339 /* The context of mbs[0]. We store the context independently, since 340 the context of mbs[0] may be different from raw_mbs[0], which is 341 the beginning of the input string. */ 342 unsigned int tip_context; 343 /* The translation passed as a part of an argument of re_compile_pattern. */ 344 RE_TRANSLATE_TYPE trans; 345 /* 1 if REG_ICASE. */ 346 unsigned int icase : 1; 347}; 348typedef struct re_string_t re_string_t; 349/* In case of REG_ICASE, we allocate the buffer dynamically for mbs. */ 350#define MBS_ALLOCATED(pstr) (pstr->icase) 351/* In case that we need translation, we allocate the buffer dynamically 352 for mbs_case. Note that mbs == mbs_case if not REG_ICASE. */ 353#define MBS_CASE_ALLOCATED(pstr) (pstr->trans != NULL) 354 355static reg_errcode_t re_string_allocate _RE_ARGS((re_string_t *pstr, const char *str, 356 int len, int init_len, 357 RE_TRANSLATE_TYPE trans, int icase)); 358static reg_errcode_t re_string_construct _RE_ARGS((re_string_t *pstr, const char *str, int len, 359 RE_TRANSLATE_TYPE trans, int icase)); 360static reg_errcode_t re_string_reconstruct _RE_ARGS((re_string_t *pstr, int idx, 361 int eflags, int newline)); 362static reg_errcode_t re_string_realloc_buffers _RE_ARGS((re_string_t *pstr, int new_buf_len)); 363#ifdef RE_ENABLE_I18N 364static void build_wcs_buffer _RE_ARGS((re_string_t *pstr)); 365static void build_wcs_upper_buffer _RE_ARGS((re_string_t *pstr)); 366#endif /* RE_ENABLE_I18N */ 367static void build_upper_buffer _RE_ARGS((re_string_t *pstr)); 368static void re_string_translate_buffer _RE_ARGS((re_string_t *pstr)); 369static void re_string_destruct _RE_ARGS((re_string_t *pstr)); 370#ifdef RE_ENABLE_I18N 371static int re_string_elem_size_at _RE_ARGS((const re_string_t *pstr, int idx)); 372static inline int re_string_char_size_at _RE_ARGS((const re_string_t *pstr, int idx)); 373static inline wint_t re_string_wchar_at _RE_ARGS((const re_string_t *pstr, int idx)); 374#endif /* RE_ENABLE_I18N */ 375static unsigned int re_string_context_at _RE_ARGS((const re_string_t *input, int idx, 376 int eflags, int newline_anchor)); 377#define re_string_peek_byte(pstr, offset) \ 378 ((pstr)->mbs[(pstr)->cur_idx + offset]) 379#define re_string_peek_byte_case(pstr, offset) \ 380 ((pstr)->mbs_case[(pstr)->cur_idx + offset]) 381#define re_string_fetch_byte(pstr) \ 382 ((pstr)->mbs[(pstr)->cur_idx++]) 383#define re_string_fetch_byte_case(pstr) \ 384 ((pstr)->mbs_case[(pstr)->cur_idx++]) 385#define re_string_first_byte(pstr, idx) \ 386 ((idx) == (pstr)->len || (pstr)->wcs[idx] != WEOF) 387#define re_string_is_single_byte_char(pstr, idx) \ 388 ((pstr)->wcs[idx] != WEOF && ((pstr)->len == (idx) \ 389 || (pstr)->wcs[(idx) + 1] != WEOF)) 390#define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx) 391#define re_string_cur_idx(pstr) ((pstr)->cur_idx) 392#define re_string_get_buffer(pstr) ((pstr)->mbs) 393#define re_string_length(pstr) ((pstr)->len) 394#define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx]) 395#define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx)) 396#define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx)) 397 398#define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t))) 399#define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t))) 400#define re_free(p) free (p) 401 402struct bin_tree_t 403{ 404 struct bin_tree_t *parent; 405 struct bin_tree_t *left; 406 struct bin_tree_t *right; 407 408 /* `node_idx' is the index in dfa->nodes, if `type' == 0. 409 Otherwise `type' indicate the type of this node. */ 410 re_token_type_t type; 411 int node_idx; 412 413 int first; 414 int next; 415 re_node_set eclosure; 416}; 417typedef struct bin_tree_t bin_tree_t; 418 419 420#define CONTEXT_WORD 1 421#define CONTEXT_NEWLINE (CONTEXT_WORD << 1) 422#define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1) 423#define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1) 424 425#define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD) 426#define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE) 427#define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF) 428#define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF) 429#define IS_ORDINARY_CONTEXT(c) ((c) == 0) 430 431#define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_') 432#define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR) 433#define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_') 434#define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR) 435 436#define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \ 437 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ 438 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ 439 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\ 440 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context))) 441 442#define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \ 443 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ 444 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ 445 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \ 446 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context))) 447 448struct re_dfastate_t 449{ 450 unsigned int hash; 451 re_node_set nodes; 452 re_node_set *entrance_nodes; 453 struct re_dfastate_t **trtable; 454 struct re_dfastate_t **trtable_search; 455 /* If this state is a special state. 456 A state is a special state if the state is the halt state, or 457 a anchor. */ 458 unsigned int context : 2; 459 unsigned int halt : 1; 460 /* If this state can accept `multi byte'. 461 Note that we refer to multibyte characters, and multi character 462 collating elements as `multi byte'. */ 463 unsigned int accept_mb : 1; 464 /* If this state has backreference node(s). */ 465 unsigned int has_backref : 1; 466 unsigned int has_constraint : 1; 467}; 468typedef struct re_dfastate_t re_dfastate_t; 469 470typedef struct 471{ 472 /* start <= node < end */ 473 int start; 474 int end; 475} re_subexp_t; 476 477struct re_state_table_entry 478{ 479 int num; 480 int alloc; 481 re_dfastate_t **array; 482}; 483 484/* Array type used in re_sub_match_last_t and re_sub_match_top_t. */ 485 486typedef struct 487{ 488 int next_idx; 489 int alloc; 490 re_dfastate_t **array; 491} state_array_t; 492 493/* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */ 494 495typedef struct 496{ 497 int node; 498 int str_idx; /* The position NODE match at. */ 499 state_array_t path; 500} re_sub_match_last_t; 501 502/* Store information about the node NODE whose type is OP_OPEN_SUBEXP. 503 And information about the node, whose type is OP_CLOSE_SUBEXP, 504 corresponding to NODE is stored in LASTS. */ 505 506typedef struct 507{ 508 int str_idx; 509 int node; 510 int next_last_offset; 511 state_array_t *path; 512 int alasts; /* Allocation size of LASTS. */ 513 int nlasts; /* The number of LASTS. */ 514 re_sub_match_last_t **lasts; 515} re_sub_match_top_t; 516 517struct re_backref_cache_entry 518{ 519 int node; 520 int str_idx; 521 int subexp_from; 522 int subexp_to; 523 int flag; 524}; 525 526typedef struct 527{ 528 /* EFLAGS of the argument of regexec. */ 529 int eflags; 530 /* Where the matching ends. */ 531 int match_last; 532 int last_node; 533 /* The string object corresponding to the input string. */ 534 re_string_t *input; 535 /* The state log used by the matcher. */ 536 re_dfastate_t **state_log; 537 int state_log_top; 538 /* Back reference cache. */ 539 int nbkref_ents; 540 int abkref_ents; 541 struct re_backref_cache_entry *bkref_ents; 542 int max_mb_elem_len; 543 int nsub_tops; 544 int asub_tops; 545 re_sub_match_top_t **sub_tops; 546} re_match_context_t; 547 548typedef struct 549{ 550 int cur_bkref; 551 int cls_subexp_idx; 552 553 re_dfastate_t **sifted_states; 554 re_dfastate_t **limited_states; 555 556 re_node_set limits; 557 558 int last_node; 559 int last_str_idx; 560 int check_subexp; 561} re_sift_context_t; 562 563struct re_fail_stack_ent_t 564{ 565 int idx; 566 int node; 567 regmatch_t *regs; 568 re_node_set eps_via_nodes; 569}; 570 571struct re_fail_stack_t 572{ 573 int num; 574 int alloc; 575 struct re_fail_stack_ent_t *stack; 576}; 577 578struct re_dfa_t 579{ 580 re_bitset_ptr_t word_char; 581 582 /* number of subexpressions `re_nsub' is in regex_t. */ 583 int subexps_alloc; 584 re_subexp_t *subexps; 585 586 re_token_t *nodes; 587 int nodes_alloc; 588 int nodes_len; 589 bin_tree_t *str_tree; 590 int *nexts; 591 int *org_indices; 592 re_node_set *edests; 593 re_node_set *eclosures; 594 re_node_set *inveclosures; 595 struct re_state_table_entry *state_table; 596 unsigned int state_hash_mask; 597 re_dfastate_t *init_state; 598 re_dfastate_t *init_state_word; 599 re_dfastate_t *init_state_nl; 600 re_dfastate_t *init_state_begbuf; 601 int states_alloc; 602 int init_node; 603 int nbackref; /* The number of backreference in this dfa. */ 604 /* Bitmap expressing which backreference is used. */ 605 unsigned int used_bkref_map; 606#ifdef DEBUG 607 char* re_str; 608#endif 609 unsigned int has_plural_match : 1; 610 /* If this dfa has "multibyte node", which is a backreference or 611 a node which can accept multibyte character or multi character 612 collating element. */ 613 unsigned int has_mb_node : 1; 614}; 615typedef struct re_dfa_t re_dfa_t; 616 617static reg_errcode_t re_node_set_alloc _RE_ARGS((re_node_set *set, int size)); 618static reg_errcode_t re_node_set_init_1 _RE_ARGS((re_node_set *set, int elem)); 619static reg_errcode_t re_node_set_init_2 _RE_ARGS((re_node_set *set, int elem1, int elem2)); 620#define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set)) 621static reg_errcode_t re_node_set_init_copy _RE_ARGS((re_node_set *dest, 622 const re_node_set *src)); 623static reg_errcode_t re_node_set_add_intersect _RE_ARGS((re_node_set *dest, 624 const re_node_set *src1, 625 const re_node_set *src2)); 626static reg_errcode_t re_node_set_init_union _RE_ARGS((re_node_set *dest, 627 const re_node_set *src1, 628 const re_node_set *src2)); 629static reg_errcode_t re_node_set_merge _RE_ARGS((re_node_set *dest, const re_node_set *src)); 630static int re_node_set_insert _RE_ARGS((re_node_set *set, int elem)); 631static int re_node_set_compare _RE_ARGS((const re_node_set *set1, const re_node_set *set2)); 632static int re_node_set_contains _RE_ARGS((const re_node_set *set, int elem)); 633static void re_node_set_remove_at _RE_ARGS((re_node_set *set, int idx)); 634#define re_node_set_remove(set,id) \ 635 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1)) 636#define re_node_set_empty(p) ((p)->nelem = 0) 637#define re_node_set_free(set) re_free ((set)->elems) 638static int re_dfa_add_node _RE_ARGS((re_dfa_t *dfa, re_token_t token, int mode)); 639static re_dfastate_t *re_acquire_state _RE_ARGS((reg_errcode_t *err, re_dfa_t *dfa, 640 const re_node_set *nodes)); 641static re_dfastate_t *re_acquire_state_context _RE_ARGS((reg_errcode_t *err, re_dfa_t *dfa, 642 const re_node_set *nodes, 643 unsigned int context)); 644static void free_state _RE_ARGS((re_dfastate_t *state)); 645 646 647typedef enum 648{ 649 SB_CHAR, 650 MB_CHAR, 651 EQUIV_CLASS, 652 COLL_SYM, 653 CHAR_CLASS 654} bracket_elem_type; 655 656typedef struct 657{ 658 bracket_elem_type type; 659 union 660 { 661 unsigned char ch; 662 unsigned char *name; 663 wchar_t wch; 664 } opr; 665} bracket_elem_t; 666 667 668/* Inline functions for bitset operation. */ 669static inline void 670bitset_not (set) 671 bitset set; 672{ 673 int bitset_i; 674 for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i) 675 set[bitset_i] = ~set[bitset_i]; 676} 677 678static inline void 679bitset_merge (dest, src) 680 bitset dest; 681 const bitset src; 682{ 683 int bitset_i; 684 for (bitset_i = 0; bitset_i < BITSET_UINTS; ++bitset_i) 685 dest[bitset_i] |= src[bitset_i]; 686} 687 688static inline void 689bitset_not_merge (dest, src) 690 bitset dest; 691 const bitset src; 692{ 693 int i; 694 for (i = 0; i < BITSET_UINTS; ++i) 695 dest[i] |= ~src[i]; 696} 697 698#ifdef RE_ENABLE_I18N 699/* Inline functions for re_string. */ 700static inline int 701re_string_char_size_at (pstr, idx) 702 const re_string_t *pstr; 703 int idx; 704{ 705 int byte_idx; 706 if (re_mb_cur_max == 1) 707 return 1; 708 for (byte_idx = 1; idx + byte_idx < pstr->len; ++byte_idx) 709 if (pstr->wcs[idx + byte_idx] != WEOF) 710 break; 711 return byte_idx; 712} 713 714static inline wint_t 715re_string_wchar_at (pstr, idx) 716 const re_string_t *pstr; 717 int idx; 718{ 719 if (re_mb_cur_max == 1) 720 return (wint_t) pstr->mbs[idx]; 721 return (wint_t) pstr->wcs[idx]; 722} 723 724static int 725re_string_elem_size_at (pstr, idx) 726 const re_string_t *pstr; 727 int idx; 728{ 729#ifdef _LIBC 730 const unsigned char *p, *extra; 731 const int32_t *table, *indirect; 732 int32_t tmp; 733# include <locale/weight.h> 734 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); 735 736 if (nrules != 0) 737 { 738 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); 739 extra = (const unsigned char *) 740 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB); 741 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE, 742 _NL_COLLATE_INDIRECTMB); 743 p = pstr->mbs + idx; 744 tmp = findidx (&p); 745 return p - pstr->mbs - idx; 746 } 747 else 748#endif /* _LIBC */ 749 return 1; 750} 751#endif /* RE_ENABLE_I18N */ 752 753#endif /* _REGEX_INTERNAL_H */ 754