1 /* Extended regular expression matching and search library. 2 Copyright (C) 2002-2005, 2007, 2008 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_LANGINFO_H || defined HAVE_LANGINFO_CODESET || defined _LIBC 31 # include <langinfo.h> 32 #endif 33 #if defined HAVE_LOCALE_H || defined _LIBC 34 # include <locale.h> 35 #endif 36 #if defined HAVE_WCHAR_H || defined _LIBC 37 # include <wchar.h> 38 #endif /* HAVE_WCHAR_H || _LIBC */ 39 #if defined HAVE_WCTYPE_H || defined _LIBC 40 # include <wctype.h> 41 #endif /* HAVE_WCTYPE_H || _LIBC */ 42 #if defined HAVE_STDBOOL_H || defined _LIBC 43 # include <stdbool.h> 44 #endif /* HAVE_STDBOOL_H || _LIBC */ 45 #if defined HAVE_STDINT_H || defined _LIBC 46 # include <stdint.h> 47 #endif /* HAVE_STDINT_H || _LIBC */ 48 #if defined _LIBC 49 # include <bits/libc-lock.h> 50 #else 51 # define __libc_lock_define(CLASS,NAME) 52 # define __libc_lock_init(NAME) do { } while (0) 53 # define __libc_lock_lock(NAME) do { } while (0) 54 # define __libc_lock_unlock(NAME) do { } while (0) 55 #endif 56 57 /* In case that the system doesn't have isblank(). */ 58 #if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank 59 # define isblank(ch) ((ch) == ' ' || (ch) == '\t') 60 #endif 61 62 #ifdef _LIBC 63 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS 64 # define _RE_DEFINE_LOCALE_FUNCTIONS 1 65 # include <locale/localeinfo.h> 66 # include <locale/elem-hash.h> 67 # include <locale/coll-lookup.h> 68 # endif 69 #endif 70 71 /* This is for other GNU distributions with internationalized messages. */ 72 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC 73 # include <libintl.h> 74 # ifdef _LIBC 75 # undef gettext 76 # define gettext(msgid) \ 77 INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES) 78 # endif 79 #else 80 # define gettext(msgid) (msgid) 81 #endif 82 83 #ifndef gettext_noop 84 /* This define is so xgettext can find the internationalizable 85 strings. */ 86 # define gettext_noop(String) String 87 #endif 88 89 /* For loser systems without the definition. */ 90 #ifndef SIZE_MAX 91 # define SIZE_MAX ((size_t) -1) 92 #endif 93 94 #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC 95 # define RE_ENABLE_I18N 96 #endif 97 98 #if __GNUC__ >= 3 99 # define BE(expr, val) __builtin_expect (expr, val) 100 #else 101 # define BE(expr, val) (expr) 102 # define inline 103 #endif 104 105 /* Number of single byte character. */ 106 #define SBC_MAX 256 107 108 #define COLL_ELEM_LEN_MAX 8 109 110 /* The character which represents newline. */ 111 #define NEWLINE_CHAR '\n' 112 #define WIDE_NEWLINE_CHAR L'\n' 113 114 /* Rename to standard API for using out of glibc. */ 115 #ifndef _LIBC 116 # define __wctype wctype 117 # define __iswctype iswctype 118 # define __btowc btowc 119 # define __mbrtowc mbrtowc 120 # define __mempcpy mempcpy 121 # define __wcrtomb wcrtomb 122 # define __regfree regfree 123 # define attribute_hidden 124 #endif /* not _LIBC */ 125 126 #ifdef __GNUC__ 127 # define __attribute(arg) __attribute__ (arg) 128 #else 129 # define __attribute(arg) 130 #endif 131 132 extern const char __re_error_msgid[] attribute_hidden; 133 extern const size_t __re_error_msgid_idx[] attribute_hidden; 134 135 /* An integer used to represent a set of bits. It must be unsigned, 136 and must be at least as wide as unsigned int. */ 137 typedef unsigned long int bitset_word_t; 138 /* All bits set in a bitset_word_t. */ 139 #define BITSET_WORD_MAX ULONG_MAX 140 /* Number of bits in a bitset_word_t. */ 141 #define BITSET_WORD_BITS (sizeof (bitset_word_t) * CHAR_BIT) 142 /* Number of bitset_word_t in a bit_set. */ 143 #define BITSET_WORDS (SBC_MAX / BITSET_WORD_BITS) 144 typedef bitset_word_t bitset_t[BITSET_WORDS]; 145 typedef bitset_word_t *re_bitset_ptr_t; 146 typedef const bitset_word_t *re_const_bitset_ptr_t; 147 148 #define bitset_set(set,i) \ 149 (set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS) 150 #define bitset_clear(set,i) \ 151 (set[i / BITSET_WORD_BITS] &= ~((bitset_word_t) 1 << i % BITSET_WORD_BITS)) 152 #define bitset_contain(set,i) \ 153 (set[i / BITSET_WORD_BITS] & ((bitset_word_t) 1 << i % BITSET_WORD_BITS)) 154 #define bitset_empty(set) memset (set, '\0', sizeof (bitset_t)) 155 #define bitset_set_all(set) memset (set, '\xff', sizeof (bitset_t)) 156 #define bitset_copy(dest,src) memcpy (dest, src, sizeof (bitset_t)) 157 158 #define PREV_WORD_CONSTRAINT 0x0001 159 #define PREV_NOTWORD_CONSTRAINT 0x0002 160 #define NEXT_WORD_CONSTRAINT 0x0004 161 #define NEXT_NOTWORD_CONSTRAINT 0x0008 162 #define PREV_NEWLINE_CONSTRAINT 0x0010 163 #define NEXT_NEWLINE_CONSTRAINT 0x0020 164 #define PREV_BEGBUF_CONSTRAINT 0x0040 165 #define NEXT_ENDBUF_CONSTRAINT 0x0080 166 #define WORD_DELIM_CONSTRAINT 0x0100 167 #define NOT_WORD_DELIM_CONSTRAINT 0x0200 168 169 typedef enum 170 { 171 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, 172 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, 173 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT, 174 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT, 175 LINE_FIRST = PREV_NEWLINE_CONSTRAINT, 176 LINE_LAST = NEXT_NEWLINE_CONSTRAINT, 177 BUF_FIRST = PREV_BEGBUF_CONSTRAINT, 178 BUF_LAST = NEXT_ENDBUF_CONSTRAINT, 179 WORD_DELIM = WORD_DELIM_CONSTRAINT, 180 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT 181 } re_context_type; 182 183 typedef struct 184 { 185 int alloc; 186 int nelem; 187 int *elems; 188 } re_node_set; 189 190 typedef enum 191 { 192 NON_TYPE = 0, 193 194 /* Node type, These are used by token, node, tree. */ 195 CHARACTER = 1, 196 END_OF_RE = 2, 197 SIMPLE_BRACKET = 3, 198 OP_BACK_REF = 4, 199 OP_PERIOD = 5, 200 #ifdef RE_ENABLE_I18N 201 COMPLEX_BRACKET = 6, 202 OP_UTF8_PERIOD = 7, 203 #endif /* RE_ENABLE_I18N */ 204 205 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used 206 when the debugger shows values of this enum type. */ 207 #define EPSILON_BIT 8 208 OP_OPEN_SUBEXP = EPSILON_BIT | 0, 209 OP_CLOSE_SUBEXP = EPSILON_BIT | 1, 210 OP_ALT = EPSILON_BIT | 2, 211 OP_DUP_ASTERISK = EPSILON_BIT | 3, 212 ANCHOR = EPSILON_BIT | 4, 213 214 /* Tree type, these are used only by tree. */ 215 CONCAT = 16, 216 SUBEXP = 17, 217 218 /* Token type, these are used only by token. */ 219 OP_DUP_PLUS = 18, 220 OP_DUP_QUESTION, 221 OP_OPEN_BRACKET, 222 OP_CLOSE_BRACKET, 223 OP_CHARSET_RANGE, 224 OP_OPEN_DUP_NUM, 225 OP_CLOSE_DUP_NUM, 226 OP_NON_MATCH_LIST, 227 OP_OPEN_COLL_ELEM, 228 OP_CLOSE_COLL_ELEM, 229 OP_OPEN_EQUIV_CLASS, 230 OP_CLOSE_EQUIV_CLASS, 231 OP_OPEN_CHAR_CLASS, 232 OP_CLOSE_CHAR_CLASS, 233 OP_WORD, 234 OP_NOTWORD, 235 OP_SPACE, 236 OP_NOTSPACE, 237 BACK_SLASH 238 239 } re_token_type_t; 240 241 #ifdef RE_ENABLE_I18N 242 typedef struct 243 { 244 /* Multibyte characters. */ 245 wchar_t *mbchars; 246 247 /* Collating symbols. */ 248 # ifdef _LIBC 249 int32_t *coll_syms; 250 # endif 251 252 /* Equivalence classes. */ 253 # ifdef _LIBC 254 int32_t *equiv_classes; 255 # endif 256 257 /* Range expressions. */ 258 # ifdef _LIBC 259 uint32_t *range_starts; 260 uint32_t *range_ends; 261 # else /* not _LIBC */ 262 wchar_t *range_starts; 263 wchar_t *range_ends; 264 # endif /* not _LIBC */ 265 266 /* Character classes. */ 267 wctype_t *char_classes; 268 269 /* If this character set is the non-matching list. */ 270 unsigned int non_match : 1; 271 272 /* # of multibyte characters. */ 273 int nmbchars; 274 275 /* # of collating symbols. */ 276 int ncoll_syms; 277 278 /* # of equivalence classes. */ 279 int nequiv_classes; 280 281 /* # of range expressions. */ 282 int nranges; 283 284 /* # of character classes. */ 285 int nchar_classes; 286 } re_charset_t; 287 #endif /* RE_ENABLE_I18N */ 288 289 typedef struct 290 { 291 union 292 { 293 unsigned char c; /* for CHARACTER */ 294 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */ 295 #ifdef RE_ENABLE_I18N 296 re_charset_t *mbcset; /* for COMPLEX_BRACKET */ 297 #endif /* RE_ENABLE_I18N */ 298 int idx; /* for BACK_REF */ 299 re_context_type ctx_type; /* for ANCHOR */ 300 } opr; 301 #if __GNUC__ >= 2 302 re_token_type_t type : 8; 303 #else 304 re_token_type_t type; 305 #endif 306 unsigned int constraint : 10; /* context constraint */ 307 unsigned int duplicated : 1; 308 unsigned int opt_subexp : 1; 309 #ifdef RE_ENABLE_I18N 310 unsigned int accept_mb : 1; 311 /* These 2 bits can be moved into the union if needed (e.g. if running out 312 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */ 313 unsigned int mb_partial : 1; 314 #endif 315 unsigned int word_char : 1; 316 } re_token_t; 317 318 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT) 319 320 struct re_string_t 321 { 322 /* Indicate the raw buffer which is the original string passed as an 323 argument of regexec(), re_search(), etc.. */ 324 const unsigned char *raw_mbs; 325 /* Store the multibyte string. In case of "case insensitive mode" like 326 REG_ICASE, upper cases of the string are stored, otherwise MBS points 327 the same address that RAW_MBS points. */ 328 unsigned char *mbs; 329 #ifdef RE_ENABLE_I18N 330 /* Store the wide character string which is corresponding to MBS. */ 331 wint_t *wcs; 332 int *offsets; 333 mbstate_t cur_state; 334 #endif 335 /* Index in RAW_MBS. Each character mbs[i] corresponds to 336 raw_mbs[raw_mbs_idx + i]. */ 337 int raw_mbs_idx; 338 /* The length of the valid characters in the buffers. */ 339 int valid_len; 340 /* The corresponding number of bytes in raw_mbs array. */ 341 int valid_raw_len; 342 /* The length of the buffers MBS and WCS. */ 343 int bufs_len; 344 /* The index in MBS, which is updated by re_string_fetch_byte. */ 345 int cur_idx; 346 /* length of RAW_MBS array. */ 347 int raw_len; 348 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */ 349 int len; 350 /* End of the buffer may be shorter than its length in the cases such 351 as re_match_2, re_search_2. Then, we use STOP for end of the buffer 352 instead of LEN. */ 353 int raw_stop; 354 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */ 355 int stop; 356 357 /* The context of mbs[0]. We store the context independently, since 358 the context of mbs[0] may be different from raw_mbs[0], which is 359 the beginning of the input string. */ 360 unsigned int tip_context; 361 /* The translation passed as a part of an argument of re_compile_pattern. */ 362 RE_TRANSLATE_TYPE trans; 363 /* Copy of re_dfa_t's word_char. */ 364 re_const_bitset_ptr_t word_char; 365 /* 1 if REG_ICASE. */ 366 unsigned char icase; 367 unsigned char is_utf8; 368 unsigned char map_notascii; 369 unsigned char mbs_allocated; 370 unsigned char offsets_needed; 371 unsigned char newline_anchor; 372 unsigned char word_ops_used; 373 int mb_cur_max; 374 }; 375 typedef struct re_string_t re_string_t; 376 377 378 struct re_dfa_t; 379 typedef struct re_dfa_t re_dfa_t; 380 381 #ifndef _LIBC 382 # ifdef __i386__ 383 # define internal_function __attribute ((regparm (3), stdcall)) 384 # else 385 # define internal_function 386 # endif 387 #endif 388 389 #ifndef NOT_IN_libc 390 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr, 391 int new_buf_len) 392 internal_function; 393 # ifdef RE_ENABLE_I18N 394 static void build_wcs_buffer (re_string_t *pstr) internal_function; 395 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr) 396 internal_function; 397 # endif /* RE_ENABLE_I18N */ 398 static void build_upper_buffer (re_string_t *pstr) internal_function; 399 static void re_string_translate_buffer (re_string_t *pstr) internal_function; 400 static unsigned int re_string_context_at (const re_string_t *input, int idx, 401 int eflags) 402 internal_function __attribute ((pure)); 403 #endif 404 #define re_string_peek_byte(pstr, offset) \ 405 ((pstr)->mbs[(pstr)->cur_idx + offset]) 406 #define re_string_fetch_byte(pstr) \ 407 ((pstr)->mbs[(pstr)->cur_idx++]) 408 #define re_string_first_byte(pstr, idx) \ 409 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF) 410 #define re_string_is_single_byte_char(pstr, idx) \ 411 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \ 412 || (pstr)->wcs[(idx) + 1] != WEOF)) 413 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx) 414 #define re_string_cur_idx(pstr) ((pstr)->cur_idx) 415 #define re_string_get_buffer(pstr) ((pstr)->mbs) 416 #define re_string_length(pstr) ((pstr)->len) 417 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx]) 418 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx)) 419 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx)) 420 421 #include <alloca.h> 422 423 #ifndef _LIBC 424 # if HAVE_ALLOCA 425 /* The OS usually guarantees only one guard page at the bottom of the stack, 426 and a page size can be as small as 4096 bytes. So we cannot safely 427 allocate anything larger than 4096 bytes. Also care for the possibility 428 of a few compiler-allocated temporary stack slots. */ 429 # define __libc_use_alloca(n) ((n) < 4032) 430 # else 431 /* alloca is implemented with malloc, so just use malloc. */ 432 # define __libc_use_alloca(n) 0 433 # endif 434 #endif 435 436 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t))) 437 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t))) 438 #define re_free(p) free (p) 439 440 struct bin_tree_t 441 { 442 struct bin_tree_t *parent; 443 struct bin_tree_t *left; 444 struct bin_tree_t *right; 445 struct bin_tree_t *first; 446 struct bin_tree_t *next; 447 448 re_token_t token; 449 450 /* `node_idx' is the index in dfa->nodes, if `type' == 0. 451 Otherwise `type' indicate the type of this node. */ 452 int node_idx; 453 }; 454 typedef struct bin_tree_t bin_tree_t; 455 456 #define BIN_TREE_STORAGE_SIZE \ 457 ((1024 - sizeof (void *)) / sizeof (bin_tree_t)) 458 459 struct bin_tree_storage_t 460 { 461 struct bin_tree_storage_t *next; 462 bin_tree_t data[BIN_TREE_STORAGE_SIZE]; 463 }; 464 typedef struct bin_tree_storage_t bin_tree_storage_t; 465 466 #define CONTEXT_WORD 1 467 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1) 468 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1) 469 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1) 470 471 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD) 472 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE) 473 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF) 474 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF) 475 #define IS_ORDINARY_CONTEXT(c) ((c) == 0) 476 477 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_') 478 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR) 479 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_') 480 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR) 481 482 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \ 483 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ 484 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ 485 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\ 486 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context))) 487 488 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \ 489 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ 490 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ 491 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \ 492 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context))) 493 494 struct re_dfastate_t 495 { 496 unsigned int hash; 497 re_node_set nodes; 498 re_node_set non_eps_nodes; 499 re_node_set inveclosure; 500 re_node_set *entrance_nodes; 501 struct re_dfastate_t **trtable, **word_trtable; 502 unsigned int context : 4; 503 unsigned int halt : 1; 504 /* If this state can accept `multi byte'. 505 Note that we refer to multibyte characters, and multi character 506 collating elements as `multi byte'. */ 507 unsigned int accept_mb : 1; 508 /* If this state has backreference node(s). */ 509 unsigned int has_backref : 1; 510 unsigned int has_constraint : 1; 511 }; 512 typedef struct re_dfastate_t re_dfastate_t; 513 514 struct re_state_table_entry 515 { 516 int num; 517 int alloc; 518 re_dfastate_t **array; 519 }; 520 521 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */ 522 523 typedef struct 524 { 525 int next_idx; 526 int alloc; 527 re_dfastate_t **array; 528 } state_array_t; 529 530 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */ 531 532 typedef struct 533 { 534 int node; 535 int str_idx; /* The position NODE match at. */ 536 state_array_t path; 537 } re_sub_match_last_t; 538 539 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP. 540 And information about the node, whose type is OP_CLOSE_SUBEXP, 541 corresponding to NODE is stored in LASTS. */ 542 543 typedef struct 544 { 545 int str_idx; 546 int node; 547 state_array_t *path; 548 int alasts; /* Allocation size of LASTS. */ 549 int nlasts; /* The number of LASTS. */ 550 re_sub_match_last_t **lasts; 551 } re_sub_match_top_t; 552 553 struct re_backref_cache_entry 554 { 555 int node; 556 int str_idx; 557 int subexp_from; 558 int subexp_to; 559 char more; 560 char unused; 561 unsigned short int eps_reachable_subexps_map; 562 }; 563 564 typedef struct 565 { 566 /* The string object corresponding to the input string. */ 567 re_string_t input; 568 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L) 569 const re_dfa_t *const dfa; 570 #else 571 const re_dfa_t *dfa; 572 #endif 573 /* EFLAGS of the argument of regexec. */ 574 int eflags; 575 /* Where the matching ends. */ 576 int match_last; 577 int last_node; 578 /* The state log used by the matcher. */ 579 re_dfastate_t **state_log; 580 int state_log_top; 581 /* Back reference cache. */ 582 int nbkref_ents; 583 int abkref_ents; 584 struct re_backref_cache_entry *bkref_ents; 585 int max_mb_elem_len; 586 int nsub_tops; 587 int asub_tops; 588 re_sub_match_top_t **sub_tops; 589 } re_match_context_t; 590 591 typedef struct 592 { 593 re_dfastate_t **sifted_states; 594 re_dfastate_t **limited_states; 595 int last_node; 596 int last_str_idx; 597 re_node_set limits; 598 } re_sift_context_t; 599 600 struct re_fail_stack_ent_t 601 { 602 int idx; 603 int node; 604 regmatch_t *regs; 605 re_node_set eps_via_nodes; 606 }; 607 608 struct re_fail_stack_t 609 { 610 int num; 611 int alloc; 612 struct re_fail_stack_ent_t *stack; 613 }; 614 615 struct re_dfa_t 616 { 617 re_token_t *nodes; 618 size_t nodes_alloc; 619 size_t nodes_len; 620 int *nexts; 621 int *org_indices; 622 re_node_set *edests; 623 re_node_set *eclosures; 624 re_node_set *inveclosures; 625 struct re_state_table_entry *state_table; 626 re_dfastate_t *init_state; 627 re_dfastate_t *init_state_word; 628 re_dfastate_t *init_state_nl; 629 re_dfastate_t *init_state_begbuf; 630 bin_tree_t *str_tree; 631 bin_tree_storage_t *str_tree_storage; 632 re_bitset_ptr_t sb_char; 633 int str_tree_storage_idx; 634 635 /* number of subexpressions `re_nsub' is in regex_t. */ 636 unsigned int state_hash_mask; 637 int init_node; 638 int nbackref; /* The number of backreference in this dfa. */ 639 640 /* Bitmap expressing which backreference is used. */ 641 bitset_word_t used_bkref_map; 642 bitset_word_t completed_bkref_map; 643 644 unsigned int has_plural_match : 1; 645 /* If this dfa has "multibyte node", which is a backreference or 646 a node which can accept multibyte character or multi character 647 collating element. */ 648 unsigned int has_mb_node : 1; 649 unsigned int is_utf8 : 1; 650 unsigned int map_notascii : 1; 651 unsigned int word_ops_used : 1; 652 int mb_cur_max; 653 bitset_t word_char; 654 reg_syntax_t syntax; 655 int *subexp_map; 656 #ifdef DEBUG 657 char* re_str; 658 #endif 659 __libc_lock_define (, lock) 660 }; 661 662 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set)) 663 #define re_node_set_remove(set,id) \ 664 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1)) 665 #define re_node_set_empty(p) ((p)->nelem = 0) 666 #define re_node_set_free(set) re_free ((set)->elems) 667 668 669 typedef enum 670 { 671 SB_CHAR, 672 MB_CHAR, 673 EQUIV_CLASS, 674 COLL_SYM, 675 CHAR_CLASS 676 } bracket_elem_type; 677 678 typedef struct 679 { 680 bracket_elem_type type; 681 union 682 { 683 unsigned char ch; 684 unsigned char *name; 685 wchar_t wch; 686 } opr; 687 } bracket_elem_t; 688 689 690 /* Inline functions for bitset operation. */ 691 static inline void 692 bitset_not (bitset_t set) 693 { 694 int bitset_i; 695 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i) 696 set[bitset_i] = ~set[bitset_i]; 697 } 698 699 static inline void 700 bitset_merge (bitset_t dest, const bitset_t src) 701 { 702 int bitset_i; 703 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i) 704 dest[bitset_i] |= src[bitset_i]; 705 } 706 707 static inline void 708 bitset_mask (bitset_t dest, const bitset_t src) 709 { 710 int bitset_i; 711 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i) 712 dest[bitset_i] &= src[bitset_i]; 713 } 714 715 #ifdef RE_ENABLE_I18N 716 /* Inline functions for re_string. */ 717 static inline int 718 internal_function __attribute ((pure)) 719 re_string_char_size_at (const re_string_t *pstr, int idx) 720 { 721 int byte_idx; 722 if (pstr->mb_cur_max == 1) 723 return 1; 724 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx) 725 if (pstr->wcs[idx + byte_idx] != WEOF) 726 break; 727 return byte_idx; 728 } 729 730 static inline wint_t 731 internal_function __attribute ((pure)) 732 re_string_wchar_at (const re_string_t *pstr, int idx) 733 { 734 if (pstr->mb_cur_max == 1) 735 return (wint_t) pstr->mbs[idx]; 736 return (wint_t) pstr->wcs[idx]; 737 } 738 739 # ifndef NOT_IN_libc 740 static int 741 internal_function __attribute ((pure)) 742 re_string_elem_size_at (const re_string_t *pstr, int idx) 743 { 744 # ifdef _LIBC 745 const unsigned char *p, *extra; 746 const int32_t *table, *indirect; 747 int32_t tmp; 748 # include <locale/weight.h> 749 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); 750 751 if (nrules != 0) 752 { 753 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); 754 extra = (const unsigned char *) 755 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB); 756 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE, 757 _NL_COLLATE_INDIRECTMB); 758 p = pstr->mbs + idx; 759 tmp = findidx (&p); 760 return p - pstr->mbs - idx; 761 } 762 else 763 # endif /* _LIBC */ 764 return 1; 765 } 766 # endif 767 #endif /* RE_ENABLE_I18N */ 768 769 #endif /* _REGEX_INTERNAL_H */ 770