1 /* Extended regular expression matching and search library. 2 Copyright (C) 2002-2013 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 General Public 8 License as published by the Free Software Foundation; either 9 version 3 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 General Public License for more details. 15 16 You should have received a copy of the GNU General Public 17 License along with the GNU C Library; if not, see 18 <http://www.gnu.org/licenses/>. */ 19 20 #ifndef _REGEX_INTERNAL_H 21 #define _REGEX_INTERNAL_H 1 22 23 #include <assert.h> 24 #include <ctype.h> 25 #include <stdio.h> 26 #include <stdlib.h> 27 #include <string.h> 28 29 #include <langinfo.h> 30 #include <locale.h> 31 #include <wchar.h> 32 #include <wctype.h> 33 #include <stdbool.h> 34 #include <stdint.h> 35 #if defined _LIBC 36 # include <bits/libc-lock.h> 37 #else 38 # define __libc_lock_init(NAME) do { } while (0) 39 # define __libc_lock_lock(NAME) do { } while (0) 40 # define __libc_lock_unlock(NAME) do { } while (0) 41 #endif 42 43 /* In case that the system doesn't have isblank(). */ 44 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK)) 45 # define isblank(ch) ((ch) == ' ' || (ch) == '\t') 46 #endif 47 48 #ifdef _LIBC 49 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS 50 # define _RE_DEFINE_LOCALE_FUNCTIONS 1 51 # include <locale/localeinfo.h> 52 # include <locale/elem-hash.h> 53 # include <locale/coll-lookup.h> 54 # endif 55 #endif 56 57 /* This is for other GNU distributions with internationalized messages. */ 58 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC 59 # include <libintl.h> 60 # ifdef _LIBC 61 # undef gettext 62 # define gettext(msgid) \ 63 __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES) 64 # endif 65 #else 66 # define gettext(msgid) (msgid) 67 #endif 68 69 #ifndef gettext_noop 70 /* This define is so xgettext can find the internationalizable 71 strings. */ 72 # define gettext_noop(String) String 73 #endif 74 75 #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE && HAVE_WCSCOLL) || _LIBC 76 # define RE_ENABLE_I18N 77 #endif 78 79 #if __GNUC__ >= 3 80 # define BE(expr, val) __builtin_expect (expr, val) 81 #else 82 # define BE(expr, val) (expr) 83 #endif 84 85 /* Number of ASCII characters. */ 86 #define ASCII_CHARS 0x80 87 88 /* Number of single byte characters. */ 89 #define SBC_MAX (UCHAR_MAX + 1) 90 91 #define COLL_ELEM_LEN_MAX 8 92 93 /* The character which represents newline. */ 94 #define NEWLINE_CHAR '\n' 95 #define WIDE_NEWLINE_CHAR L'\n' 96 97 /* Rename to standard API for using out of glibc. */ 98 #ifndef _LIBC 99 # undef __wctype 100 # undef __iswctype 101 # define __wctype wctype 102 # define __iswctype iswctype 103 # define __btowc btowc 104 # define __mbrtowc mbrtowc 105 # define __wcrtomb wcrtomb 106 # define __regfree regfree 107 # define attribute_hidden 108 #endif /* not _LIBC */ 109 110 #if __GNUC__ < 3 + (__GNUC_MINOR__ < 1) 111 # define __attribute__(arg) 112 #endif 113 114 typedef __re_idx_t Idx; 115 #ifdef _REGEX_LARGE_OFFSETS 116 # define IDX_MAX (SIZE_MAX - 2) 117 #else 118 # define IDX_MAX INT_MAX 119 #endif 120 121 /* Special return value for failure to match. */ 122 #define REG_MISSING ((Idx) -1) 123 124 /* Special return value for internal error. */ 125 #define REG_ERROR ((Idx) -2) 126 127 /* Test whether N is a valid index, and is not one of the above. */ 128 #ifdef _REGEX_LARGE_OFFSETS 129 # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR) 130 #else 131 # define REG_VALID_INDEX(n) (0 <= (n)) 132 #endif 133 134 /* Test whether N is a valid nonzero index. */ 135 #ifdef _REGEX_LARGE_OFFSETS 136 # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1)) 137 #else 138 # define REG_VALID_NONZERO_INDEX(n) (0 < (n)) 139 #endif 140 141 /* A hash value, suitable for computing hash tables. */ 142 typedef __re_size_t re_hashval_t; 143 144 /* An integer used to represent a set of bits. It must be unsigned, 145 and must be at least as wide as unsigned int. */ 146 typedef unsigned long int bitset_word_t; 147 /* All bits set in a bitset_word_t. */ 148 #define BITSET_WORD_MAX ULONG_MAX 149 150 /* Number of bits in a bitset_word_t. For portability to hosts with 151 padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)'; 152 instead, deduce it directly from BITSET_WORD_MAX. Avoid 153 greater-than-32-bit integers and unconditional shifts by more than 154 31 bits, as they're not portable. */ 155 #if BITSET_WORD_MAX == 0xffffffffUL 156 # define BITSET_WORD_BITS 32 157 #elif BITSET_WORD_MAX >> 31 >> 4 == 1 158 # define BITSET_WORD_BITS 36 159 #elif BITSET_WORD_MAX >> 31 >> 16 == 1 160 # define BITSET_WORD_BITS 48 161 #elif BITSET_WORD_MAX >> 31 >> 28 == 1 162 # define BITSET_WORD_BITS 60 163 #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1 164 # define BITSET_WORD_BITS 64 165 #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1 166 # define BITSET_WORD_BITS 72 167 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1 168 # define BITSET_WORD_BITS 128 169 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1 170 # define BITSET_WORD_BITS 256 171 #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1 172 # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */ 173 # if BITSET_WORD_BITS <= SBC_MAX 174 # error "Invalid SBC_MAX" 175 # endif 176 #else 177 # error "Add case for new bitset_word_t size" 178 #endif 179 180 /* Number of bitset_word_t values in a bitset_t. */ 181 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS) 182 183 typedef bitset_word_t bitset_t[BITSET_WORDS]; 184 typedef bitset_word_t *re_bitset_ptr_t; 185 typedef const bitset_word_t *re_const_bitset_ptr_t; 186 187 #define PREV_WORD_CONSTRAINT 0x0001 188 #define PREV_NOTWORD_CONSTRAINT 0x0002 189 #define NEXT_WORD_CONSTRAINT 0x0004 190 #define NEXT_NOTWORD_CONSTRAINT 0x0008 191 #define PREV_NEWLINE_CONSTRAINT 0x0010 192 #define NEXT_NEWLINE_CONSTRAINT 0x0020 193 #define PREV_BEGBUF_CONSTRAINT 0x0040 194 #define NEXT_ENDBUF_CONSTRAINT 0x0080 195 #define WORD_DELIM_CONSTRAINT 0x0100 196 #define NOT_WORD_DELIM_CONSTRAINT 0x0200 197 198 typedef enum 199 { 200 INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, 201 WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, 202 WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT, 203 INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT, 204 LINE_FIRST = PREV_NEWLINE_CONSTRAINT, 205 LINE_LAST = NEXT_NEWLINE_CONSTRAINT, 206 BUF_FIRST = PREV_BEGBUF_CONSTRAINT, 207 BUF_LAST = NEXT_ENDBUF_CONSTRAINT, 208 WORD_DELIM = WORD_DELIM_CONSTRAINT, 209 NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT 210 } re_context_type; 211 212 typedef struct 213 { 214 Idx alloc; 215 Idx nelem; 216 Idx *elems; 217 } re_node_set; 218 219 typedef enum 220 { 221 NON_TYPE = 0, 222 223 /* Node type, These are used by token, node, tree. */ 224 CHARACTER = 1, 225 END_OF_RE = 2, 226 SIMPLE_BRACKET = 3, 227 OP_BACK_REF = 4, 228 OP_PERIOD = 5, 229 #ifdef RE_ENABLE_I18N 230 COMPLEX_BRACKET = 6, 231 OP_UTF8_PERIOD = 7, 232 #endif /* RE_ENABLE_I18N */ 233 234 /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used 235 when the debugger shows values of this enum type. */ 236 #define EPSILON_BIT 8 237 OP_OPEN_SUBEXP = EPSILON_BIT | 0, 238 OP_CLOSE_SUBEXP = EPSILON_BIT | 1, 239 OP_ALT = EPSILON_BIT | 2, 240 OP_DUP_ASTERISK = EPSILON_BIT | 3, 241 ANCHOR = EPSILON_BIT | 4, 242 243 /* Tree type, these are used only by tree. */ 244 CONCAT = 16, 245 SUBEXP = 17, 246 247 /* Token type, these are used only by token. */ 248 OP_DUP_PLUS = 18, 249 OP_DUP_QUESTION, 250 OP_OPEN_BRACKET, 251 OP_CLOSE_BRACKET, 252 OP_CHARSET_RANGE, 253 OP_OPEN_DUP_NUM, 254 OP_CLOSE_DUP_NUM, 255 OP_NON_MATCH_LIST, 256 OP_OPEN_COLL_ELEM, 257 OP_CLOSE_COLL_ELEM, 258 OP_OPEN_EQUIV_CLASS, 259 OP_CLOSE_EQUIV_CLASS, 260 OP_OPEN_CHAR_CLASS, 261 OP_CLOSE_CHAR_CLASS, 262 OP_WORD, 263 OP_NOTWORD, 264 OP_SPACE, 265 OP_NOTSPACE, 266 BACK_SLASH 267 268 } re_token_type_t; 269 270 #ifdef RE_ENABLE_I18N 271 typedef struct 272 { 273 /* Multibyte characters. */ 274 wchar_t *mbchars; 275 276 /* Collating symbols. */ 277 # ifdef _LIBC 278 int32_t *coll_syms; 279 # endif 280 281 /* Equivalence classes. */ 282 # ifdef _LIBC 283 int32_t *equiv_classes; 284 # endif 285 286 /* Range expressions. */ 287 # ifdef _LIBC 288 uint32_t *range_starts; 289 uint32_t *range_ends; 290 # else /* not _LIBC */ 291 wchar_t *range_starts; 292 wchar_t *range_ends; 293 # endif /* not _LIBC */ 294 295 /* Character classes. */ 296 wctype_t *char_classes; 297 298 /* If this character set is the non-matching list. */ 299 unsigned int non_match : 1; 300 301 /* # of multibyte characters. */ 302 Idx nmbchars; 303 304 /* # of collating symbols. */ 305 Idx ncoll_syms; 306 307 /* # of equivalence classes. */ 308 Idx nequiv_classes; 309 310 /* # of range expressions. */ 311 Idx nranges; 312 313 /* # of character classes. */ 314 Idx nchar_classes; 315 } re_charset_t; 316 #endif /* RE_ENABLE_I18N */ 317 318 typedef struct 319 { 320 union 321 { 322 unsigned char c; /* for CHARACTER */ 323 re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */ 324 #ifdef RE_ENABLE_I18N 325 re_charset_t *mbcset; /* for COMPLEX_BRACKET */ 326 #endif /* RE_ENABLE_I18N */ 327 Idx idx; /* for BACK_REF */ 328 re_context_type ctx_type; /* for ANCHOR */ 329 } opr; 330 #if __GNUC__ >= 2 && !defined __STRICT_ANSI__ 331 re_token_type_t type : 8; 332 #else 333 re_token_type_t type; 334 #endif 335 unsigned int constraint : 10; /* context constraint */ 336 unsigned int duplicated : 1; 337 unsigned int opt_subexp : 1; 338 #ifdef RE_ENABLE_I18N 339 unsigned int accept_mb : 1; 340 /* These 2 bits can be moved into the union if needed (e.g. if running out 341 of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */ 342 unsigned int mb_partial : 1; 343 #endif 344 unsigned int word_char : 1; 345 } re_token_t; 346 347 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT) 348 349 struct re_string_t 350 { 351 /* Indicate the raw buffer which is the original string passed as an 352 argument of regexec(), re_search(), etc.. */ 353 const unsigned char *raw_mbs; 354 /* Store the multibyte string. In case of "case insensitive mode" like 355 REG_ICASE, upper cases of the string are stored, otherwise MBS points 356 the same address that RAW_MBS points. */ 357 unsigned char *mbs; 358 #ifdef RE_ENABLE_I18N 359 /* Store the wide character string which is corresponding to MBS. */ 360 wint_t *wcs; 361 Idx *offsets; 362 mbstate_t cur_state; 363 #endif 364 /* Index in RAW_MBS. Each character mbs[i] corresponds to 365 raw_mbs[raw_mbs_idx + i]. */ 366 Idx raw_mbs_idx; 367 /* The length of the valid characters in the buffers. */ 368 Idx valid_len; 369 /* The corresponding number of bytes in raw_mbs array. */ 370 Idx valid_raw_len; 371 /* The length of the buffers MBS and WCS. */ 372 Idx bufs_len; 373 /* The index in MBS, which is updated by re_string_fetch_byte. */ 374 Idx cur_idx; 375 /* length of RAW_MBS array. */ 376 Idx raw_len; 377 /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */ 378 Idx len; 379 /* End of the buffer may be shorter than its length in the cases such 380 as re_match_2, re_search_2. Then, we use STOP for end of the buffer 381 instead of LEN. */ 382 Idx raw_stop; 383 /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */ 384 Idx stop; 385 386 /* The context of mbs[0]. We store the context independently, since 387 the context of mbs[0] may be different from raw_mbs[0], which is 388 the beginning of the input string. */ 389 unsigned int tip_context; 390 /* The translation passed as a part of an argument of re_compile_pattern. */ 391 RE_TRANSLATE_TYPE trans; 392 /* Copy of re_dfa_t's word_char. */ 393 re_const_bitset_ptr_t word_char; 394 /* true if REG_ICASE. */ 395 unsigned char icase; 396 unsigned char is_utf8; 397 unsigned char map_notascii; 398 unsigned char mbs_allocated; 399 unsigned char offsets_needed; 400 unsigned char newline_anchor; 401 unsigned char word_ops_used; 402 int mb_cur_max; 403 }; 404 typedef struct re_string_t re_string_t; 405 406 407 struct re_dfa_t; 408 typedef struct re_dfa_t re_dfa_t; 409 410 #ifndef _LIBC 411 # define internal_function 412 #endif 413 414 #ifndef NOT_IN_libc 415 static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr, 416 Idx new_buf_len) 417 internal_function; 418 # ifdef RE_ENABLE_I18N 419 static void build_wcs_buffer (re_string_t *pstr) internal_function; 420 static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr) 421 internal_function; 422 # endif /* RE_ENABLE_I18N */ 423 static void build_upper_buffer (re_string_t *pstr) internal_function; 424 static void re_string_translate_buffer (re_string_t *pstr) internal_function; 425 static unsigned int re_string_context_at (const re_string_t *input, Idx idx, 426 int eflags) 427 internal_function __attribute__ ((pure)); 428 #endif 429 #define re_string_peek_byte(pstr, offset) \ 430 ((pstr)->mbs[(pstr)->cur_idx + offset]) 431 #define re_string_fetch_byte(pstr) \ 432 ((pstr)->mbs[(pstr)->cur_idx++]) 433 #define re_string_first_byte(pstr, idx) \ 434 ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF) 435 #define re_string_is_single_byte_char(pstr, idx) \ 436 ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \ 437 || (pstr)->wcs[(idx) + 1] != WEOF)) 438 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx) 439 #define re_string_cur_idx(pstr) ((pstr)->cur_idx) 440 #define re_string_get_buffer(pstr) ((pstr)->mbs) 441 #define re_string_length(pstr) ((pstr)->len) 442 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx]) 443 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx)) 444 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx)) 445 446 #if defined _LIBC || HAVE_ALLOCA 447 # include <alloca.h> 448 #endif 449 450 #ifndef _LIBC 451 # if HAVE_ALLOCA 452 /* The OS usually guarantees only one guard page at the bottom of the stack, 453 and a page size can be as small as 4096 bytes. So we cannot safely 454 allocate anything larger than 4096 bytes. Also care for the possibility 455 of a few compiler-allocated temporary stack slots. */ 456 # define __libc_use_alloca(n) ((n) < 4032) 457 # else 458 /* alloca is implemented with malloc, so just use malloc. */ 459 # define __libc_use_alloca(n) 0 460 # undef alloca 461 # define alloca(n) malloc (n) 462 # endif 463 #endif 464 465 #ifdef _LIBC 466 # define MALLOC_0_IS_NONNULL 1 467 #elif !defined MALLOC_0_IS_NONNULL 468 # define MALLOC_0_IS_NONNULL 0 469 #endif 470 471 #ifndef MAX 472 # define MAX(a,b) ((a) < (b) ? (b) : (a)) 473 #endif 474 #ifndef MIN 475 # define MIN(a,b) ((a) < (b) ? (a) : (b)) 476 #endif 477 478 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t))) 479 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t))) 480 #define re_free(p) free (p) 481 482 struct bin_tree_t 483 { 484 struct bin_tree_t *parent; 485 struct bin_tree_t *left; 486 struct bin_tree_t *right; 487 struct bin_tree_t *first; 488 struct bin_tree_t *next; 489 490 re_token_t token; 491 492 /* 'node_idx' is the index in dfa->nodes, if 'type' == 0. 493 Otherwise 'type' indicate the type of this node. */ 494 Idx node_idx; 495 }; 496 typedef struct bin_tree_t bin_tree_t; 497 498 #define BIN_TREE_STORAGE_SIZE \ 499 ((1024 - sizeof (void *)) / sizeof (bin_tree_t)) 500 501 struct bin_tree_storage_t 502 { 503 struct bin_tree_storage_t *next; 504 bin_tree_t data[BIN_TREE_STORAGE_SIZE]; 505 }; 506 typedef struct bin_tree_storage_t bin_tree_storage_t; 507 508 #define CONTEXT_WORD 1 509 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1) 510 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1) 511 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1) 512 513 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD) 514 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE) 515 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF) 516 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF) 517 #define IS_ORDINARY_CONTEXT(c) ((c) == 0) 518 519 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_') 520 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR) 521 #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_') 522 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR) 523 524 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \ 525 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ 526 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ 527 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\ 528 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context))) 529 530 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \ 531 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ 532 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ 533 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \ 534 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context))) 535 536 struct re_dfastate_t 537 { 538 re_hashval_t hash; 539 re_node_set nodes; 540 re_node_set non_eps_nodes; 541 re_node_set inveclosure; 542 re_node_set *entrance_nodes; 543 struct re_dfastate_t **trtable, **word_trtable; 544 unsigned int context : 4; 545 unsigned int halt : 1; 546 /* If this state can accept "multi byte". 547 Note that we refer to multibyte characters, and multi character 548 collating elements as "multi byte". */ 549 unsigned int accept_mb : 1; 550 /* If this state has backreference node(s). */ 551 unsigned int has_backref : 1; 552 unsigned int has_constraint : 1; 553 }; 554 typedef struct re_dfastate_t re_dfastate_t; 555 556 struct re_state_table_entry 557 { 558 Idx num; 559 Idx alloc; 560 re_dfastate_t **array; 561 }; 562 563 /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */ 564 565 typedef struct 566 { 567 Idx next_idx; 568 Idx alloc; 569 re_dfastate_t **array; 570 } state_array_t; 571 572 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */ 573 574 typedef struct 575 { 576 Idx node; 577 Idx str_idx; /* The position NODE match at. */ 578 state_array_t path; 579 } re_sub_match_last_t; 580 581 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP. 582 And information about the node, whose type is OP_CLOSE_SUBEXP, 583 corresponding to NODE is stored in LASTS. */ 584 585 typedef struct 586 { 587 Idx str_idx; 588 Idx node; 589 state_array_t *path; 590 Idx alasts; /* Allocation size of LASTS. */ 591 Idx nlasts; /* The number of LASTS. */ 592 re_sub_match_last_t **lasts; 593 } re_sub_match_top_t; 594 595 struct re_backref_cache_entry 596 { 597 Idx node; 598 Idx str_idx; 599 Idx subexp_from; 600 Idx subexp_to; 601 char more; 602 char unused; 603 unsigned short int eps_reachable_subexps_map; 604 }; 605 606 typedef struct 607 { 608 /* The string object corresponding to the input string. */ 609 re_string_t input; 610 #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L) 611 const re_dfa_t *const dfa; 612 #else 613 const re_dfa_t *dfa; 614 #endif 615 /* EFLAGS of the argument of regexec. */ 616 int eflags; 617 /* Where the matching ends. */ 618 Idx match_last; 619 Idx last_node; 620 /* The state log used by the matcher. */ 621 re_dfastate_t **state_log; 622 Idx state_log_top; 623 /* Back reference cache. */ 624 Idx nbkref_ents; 625 Idx abkref_ents; 626 struct re_backref_cache_entry *bkref_ents; 627 int max_mb_elem_len; 628 Idx nsub_tops; 629 Idx asub_tops; 630 re_sub_match_top_t **sub_tops; 631 } re_match_context_t; 632 633 typedef struct 634 { 635 re_dfastate_t **sifted_states; 636 re_dfastate_t **limited_states; 637 Idx last_node; 638 Idx last_str_idx; 639 re_node_set limits; 640 } re_sift_context_t; 641 642 struct re_fail_stack_ent_t 643 { 644 Idx idx; 645 Idx node; 646 regmatch_t *regs; 647 re_node_set eps_via_nodes; 648 }; 649 650 struct re_fail_stack_t 651 { 652 Idx num; 653 Idx alloc; 654 struct re_fail_stack_ent_t *stack; 655 }; 656 657 struct re_dfa_t 658 { 659 re_token_t *nodes; 660 size_t nodes_alloc; 661 size_t nodes_len; 662 Idx *nexts; 663 Idx *org_indices; 664 re_node_set *edests; 665 re_node_set *eclosures; 666 re_node_set *inveclosures; 667 struct re_state_table_entry *state_table; 668 re_dfastate_t *init_state; 669 re_dfastate_t *init_state_word; 670 re_dfastate_t *init_state_nl; 671 re_dfastate_t *init_state_begbuf; 672 bin_tree_t *str_tree; 673 bin_tree_storage_t *str_tree_storage; 674 re_bitset_ptr_t sb_char; 675 int str_tree_storage_idx; 676 677 /* number of subexpressions 're_nsub' is in regex_t. */ 678 re_hashval_t state_hash_mask; 679 Idx init_node; 680 Idx nbackref; /* The number of backreference in this dfa. */ 681 682 /* Bitmap expressing which backreference is used. */ 683 bitset_word_t used_bkref_map; 684 bitset_word_t completed_bkref_map; 685 686 unsigned int has_plural_match : 1; 687 /* If this dfa has "multibyte node", which is a backreference or 688 a node which can accept multibyte character or multi character 689 collating element. */ 690 unsigned int has_mb_node : 1; 691 unsigned int is_utf8 : 1; 692 unsigned int map_notascii : 1; 693 unsigned int word_ops_used : 1; 694 int mb_cur_max; 695 bitset_t word_char; 696 reg_syntax_t syntax; 697 Idx *subexp_map; 698 #ifdef DEBUG 699 char* re_str; 700 #endif 701 #ifdef _LIBC 702 __libc_lock_define (, lock) 703 #endif 704 }; 705 706 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set)) 707 #define re_node_set_remove(set,id) \ 708 (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1)) 709 #define re_node_set_empty(p) ((p)->nelem = 0) 710 #define re_node_set_free(set) re_free ((set)->elems) 711 712 713 typedef enum 714 { 715 SB_CHAR, 716 MB_CHAR, 717 EQUIV_CLASS, 718 COLL_SYM, 719 CHAR_CLASS 720 } bracket_elem_type; 721 722 typedef struct 723 { 724 bracket_elem_type type; 725 union 726 { 727 unsigned char ch; 728 unsigned char *name; 729 wchar_t wch; 730 } opr; 731 } bracket_elem_t; 732 733 734 /* Functions for bitset_t operation. */ 735 736 static void 737 bitset_set (bitset_t set, Idx i) 738 { 739 set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS; 740 } 741 742 static void 743 bitset_clear (bitset_t set, Idx i) 744 { 745 set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS); 746 } 747 748 static bool 749 bitset_contain (const bitset_t set, Idx i) 750 { 751 return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1; 752 } 753 754 static void 755 bitset_empty (bitset_t set) 756 { 757 memset (set, '\0', sizeof (bitset_t)); 758 } 759 760 static void 761 bitset_set_all (bitset_t set) 762 { 763 memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS)); 764 if (SBC_MAX % BITSET_WORD_BITS != 0) 765 set[BITSET_WORDS - 1] = 766 ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1; 767 } 768 769 static void 770 bitset_copy (bitset_t dest, const bitset_t src) 771 { 772 memcpy (dest, src, sizeof (bitset_t)); 773 } 774 775 static void __attribute__ ((unused)) 776 bitset_not (bitset_t set) 777 { 778 int bitset_i; 779 for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i) 780 set[bitset_i] = ~set[bitset_i]; 781 if (SBC_MAX % BITSET_WORD_BITS != 0) 782 set[BITSET_WORDS - 1] = 783 ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1) 784 & ~set[BITSET_WORDS - 1]); 785 } 786 787 static void __attribute__ ((unused)) 788 bitset_merge (bitset_t dest, const bitset_t src) 789 { 790 int bitset_i; 791 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i) 792 dest[bitset_i] |= src[bitset_i]; 793 } 794 795 static void __attribute__ ((unused)) 796 bitset_mask (bitset_t dest, const bitset_t src) 797 { 798 int bitset_i; 799 for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i) 800 dest[bitset_i] &= src[bitset_i]; 801 } 802 803 #ifdef RE_ENABLE_I18N 804 /* Functions for re_string. */ 805 static int 806 internal_function __attribute__ ((pure, unused)) 807 re_string_char_size_at (const re_string_t *pstr, Idx idx) 808 { 809 int byte_idx; 810 if (pstr->mb_cur_max == 1) 811 return 1; 812 for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx) 813 if (pstr->wcs[idx + byte_idx] != WEOF) 814 break; 815 return byte_idx; 816 } 817 818 static wint_t 819 internal_function __attribute__ ((pure, unused)) 820 re_string_wchar_at (const re_string_t *pstr, Idx idx) 821 { 822 if (pstr->mb_cur_max == 1) 823 return (wint_t) pstr->mbs[idx]; 824 return (wint_t) pstr->wcs[idx]; 825 } 826 827 # ifndef NOT_IN_libc 828 static int 829 internal_function __attribute__ ((pure, unused)) 830 re_string_elem_size_at (const re_string_t *pstr, Idx idx) 831 { 832 # ifdef _LIBC 833 const unsigned char *p, *extra; 834 const int32_t *table, *indirect; 835 # include <locale/weight.h> 836 uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); 837 838 if (nrules != 0) 839 { 840 table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); 841 extra = (const unsigned char *) 842 _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB); 843 indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE, 844 _NL_COLLATE_INDIRECTMB); 845 p = pstr->mbs + idx; 846 findidx (&p, pstr->len - idx); 847 return p - pstr->mbs - idx; 848 } 849 else 850 # endif /* _LIBC */ 851 return 1; 852 } 853 # endif 854 #endif /* RE_ENABLE_I18N */ 855 856 #ifndef __GNUC_PREREQ 857 # if defined __GNUC__ && defined __GNUC_MINOR__ 858 # define __GNUC_PREREQ(maj, min) \ 859 ((__GNUC__ << 16) + __GNUC_MINOR__ >= ((maj) << 16) + (min)) 860 # else 861 # define __GNUC_PREREQ(maj, min) 0 862 # endif 863 #endif 864 865 #if __GNUC_PREREQ (3,4) 866 # undef __attribute_warn_unused_result__ 867 # define __attribute_warn_unused_result__ \ 868 __attribute__ ((__warn_unused_result__)) 869 #else 870 # define __attribute_warn_unused_result__ /* empty */ 871 #endif 872 873 #endif /* _REGEX_INTERNAL_H */ 874