posix/regex_internal.h

   1 /* Extended regular expression matching and search library.
   2    Copyright (C) 2002-2021 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, see
  18    <https://www.gnu.org/licenses/>.  */
  19
  20 #ifndef _REGEX_INTERNAL_H
  21 #define _REGEX_INTERNAL_H 1
  22
  23 #include <ctype.h>
  24 #include <stdio.h>
  25 #include <stdlib.h>
  26 #include <string.h>
  27
  28 #include <langinfo.h>
  29 #include <locale.h>
  30 #include <wchar.h>
  31 #include <wctype.h>
  32 #include <stdbool.h>
  33 #include <stdint.h>
  34
  35 #include <intprops.h>
  36 #include <verify.h>
  37
  38 #if defined DEBUG && DEBUG != 0
  39 # include <assert.h>
  40 # define DEBUG_ASSERT(x) assert (x)
  41 #else
  42 # define DEBUG_ASSERT(x) assume (x)
  43 #endif
  44
  45 #ifdef _LIBC
  46 # include <libc-lock.h>
  47 # define lock_define(name) __libc_lock_define (, name)
  48 # define lock_init(lock) (__libc_lock_init (lock), 0)
  49 # define lock_fini(lock) ((void) 0)
  50 # define lock_lock(lock) __libc_lock_lock (lock)
  51 # define lock_unlock(lock) __libc_lock_unlock (lock)
  52 #elif defined GNULIB_LOCK && !defined USE_UNLOCKED_IO
  53 # include "glthread/lock.h"
  54 # define lock_define(name) gl_lock_define (, name)
  55 # define lock_init(lock) glthread_lock_init (&(lock))
  56 # define lock_fini(lock) glthread_lock_destroy (&(lock))
  57 # define lock_lock(lock) glthread_lock_lock (&(lock))
  58 # define lock_unlock(lock) glthread_lock_unlock (&(lock))
  59 #elif defined GNULIB_PTHREAD && !defined USE_UNLOCKED_IO
  60 # include <pthread.h>
  61 # define lock_define(name) pthread_mutex_t name;
  62 # define lock_init(lock) pthread_mutex_init (&(lock), 0)
  63 # define lock_fini(lock) pthread_mutex_destroy (&(lock))
  64 # define lock_lock(lock) pthread_mutex_lock (&(lock))
  65 # define lock_unlock(lock) pthread_mutex_unlock (&(lock))
  66 #else
  67 # define lock_define(name)
  68 # define lock_init(lock) 0
  69 # define lock_fini(lock) ((void) 0)
  70   /* The 'dfa' avoids an "unused variable 'dfa'" warning from GCC.  */
  71 # define lock_lock(lock) ((void) dfa)
  72 # define lock_unlock(lock) ((void) 0)
  73 #endif
  74
  75 /* In case that the system doesn't have isblank().  */
  76 #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
  77 # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
  78 #endif
  79
  80 #ifdef _LIBC
  81 # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
  82 #  define _RE_DEFINE_LOCALE_FUNCTIONS 1
  83 #   include <locale/localeinfo.h>
  84 #   include <locale/coll-lookup.h>
  85 # endif
  86 #endif
  87
  88 /* This is for other GNU distributions with internationalized messages.  */
  89 #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
  90 # include <libintl.h>
  91 # ifdef _LIBC
  92 #  undef gettext
  93 #  define gettext(msgid) \
  94   __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
  95 # endif
  96 #else
  97 # undef gettext
  98 # define gettext(msgid) (msgid)
  99 #endif
 100
 101 #ifndef gettext_noop
 102 /* This define is so xgettext can find the internationalizable
 103    strings.  */
 104 # define gettext_noop(String) String
 105 #endif
 106
 107 #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE) || _LIBC
 108 # define RE_ENABLE_I18N
 109 #endif
 110
 111 /* Number of ASCII characters.  */
 112 #define ASCII_CHARS 0x80
 113
 114 /* Number of single byte characters.  */
 115 #define SBC_MAX (UCHAR_MAX + 1)
 116
 117 #define COLL_ELEM_LEN_MAX 8
 118
 119 /* The character which represents newline.  */
 120 #define NEWLINE_CHAR '\n'
 121 #define WIDE_NEWLINE_CHAR L'\n'
 122
 123 /* Rename to standard API for using out of glibc.  */
 124 #ifndef _LIBC
 125 # undef __wctype
 126 # undef __iswalnum
 127 # undef __iswctype
 128 # undef __towlower
 129 # undef __towupper
 130 # define __wctype wctype
 131 # define __iswalnum iswalnum
 132 # define __iswctype iswctype
 133 # define __towlower towlower
 134 # define __towupper towupper
 135 # define __btowc btowc
 136 # define __mbrtowc mbrtowc
 137 # define __wcrtomb wcrtomb
 138 # define __regfree regfree
 139 #endif /* not _LIBC */
 140
 141 #ifndef SSIZE_MAX
 142 # define SSIZE_MAX ((ssize_t) (SIZE_MAX / 2))
 143 #endif
 144 #ifndef ULONG_WIDTH
 145 # define ULONG_WIDTH REGEX_UINTEGER_WIDTH (ULONG_MAX)
 146 /* The number of usable bits in an unsigned integer type with maximum
 147    value MAX, as an int expression suitable in #if.  Cover all known
 148    practical hosts.  This implementation exploits the fact that MAX is
 149    1 less than a power of 2, and merely counts the number of 1 bits in
 150    MAX; "COBn" means "count the number of 1 bits in the low-order n bits".  */
 151 # define REGEX_UINTEGER_WIDTH(max) REGEX_COB128 (max)
 152 # define REGEX_COB128(n) (REGEX_COB64 ((n) >> 31 >> 31 >> 2) + REGEX_COB64 (n))
 153 # define REGEX_COB64(n) (REGEX_COB32 ((n) >> 31 >> 1) + REGEX_COB32 (n))
 154 # define REGEX_COB32(n) (REGEX_COB16 ((n) >> 16) + REGEX_COB16 (n))
 155 # define REGEX_COB16(n) (REGEX_COB8 ((n) >> 8) + REGEX_COB8 (n))
 156 # define REGEX_COB8(n) (REGEX_COB4 ((n) >> 4) + REGEX_COB4 (n))
 157 # define REGEX_COB4(n) (!!((n) & 8) + !!((n) & 4) + !!((n) & 2) + ((n) & 1))
 158 # if ULONG_MAX / 2 + 1 != 1ul << (ULONG_WIDTH - 1)
 159 #  error "ULONG_MAX out of range"
 160 # endif
 161 #endif
 162
 163 /* The type of indexes into strings.  This is signed, not size_t,
 164    since the API requires indexes to fit in regoff_t anyway, and using
 165    signed integers makes the code a bit smaller and presumably faster.
 166    The traditional GNU regex implementation uses int for indexes.
 167    The POSIX-compatible implementation uses a possibly-wider type.
 168    The name 'Idx' is three letters to minimize the hassle of
 169    reindenting a lot of regex code that formerly used 'int'.  */
 170 typedef regoff_t Idx;
 171 #ifdef _REGEX_LARGE_OFFSETS
 172 # define IDX_MAX SSIZE_MAX
 173 #else
 174 # define IDX_MAX INT_MAX
 175 #endif
 176
 177 /* A hash value, suitable for computing hash tables.  */
 178 typedef __re_size_t re_hashval_t;
 179
 180 /* An integer used to represent a set of bits.  It must be unsigned,
 181    and must be at least as wide as unsigned int.  */
 182 typedef unsigned long int bitset_word_t;
 183 /* All bits set in a bitset_word_t.  */
 184 #define BITSET_WORD_MAX ULONG_MAX
 185 /* Number of bits in a bitset_word_t.  */
 186 #define BITSET_WORD_BITS ULONG_WIDTH
 187
 188 /* Number of bitset_word_t values in a bitset_t.  */
 189 #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
 190
 191 typedef bitset_word_t bitset_t[BITSET_WORDS];
 192 typedef bitset_word_t *re_bitset_ptr_t;
 193 typedef const bitset_word_t *re_const_bitset_ptr_t;
 194
 195 #define PREV_WORD_CONSTRAINT 0x0001
 196 #define PREV_NOTWORD_CONSTRAINT 0x0002
 197 #define NEXT_WORD_CONSTRAINT 0x0004
 198 #define NEXT_NOTWORD_CONSTRAINT 0x0008
 199 #define PREV_NEWLINE_CONSTRAINT 0x0010
 200 #define NEXT_NEWLINE_CONSTRAINT 0x0020
 201 #define PREV_BEGBUF_CONSTRAINT 0x0040
 202 #define NEXT_ENDBUF_CONSTRAINT 0x0080
 203 #define WORD_DELIM_CONSTRAINT 0x0100
 204 #define NOT_WORD_DELIM_CONSTRAINT 0x0200
 205
 206 typedef enum
 207 {
 208   INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
 209   WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
 210   WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
 211   INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
 212   LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
 213   LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
 214   BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
 215   BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
 216   WORD_DELIM = WORD_DELIM_CONSTRAINT,
 217   NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
 218 } re_context_type;
 219
 220 typedef struct
 221 {
 222   Idx alloc;
 223   Idx nelem;
 224   Idx *elems;
 225 } re_node_set;
 226
 227 typedef enum
 228 {
 229   NON_TYPE = 0,
 230
 231   /* Node type, These are used by token, node, tree.  */
 232   CHARACTER = 1,
 233   END_OF_RE = 2,
 234   SIMPLE_BRACKET = 3,
 235   OP_BACK_REF = 4,
 236   OP_PERIOD = 5,
 237 #ifdef RE_ENABLE_I18N
 238   COMPLEX_BRACKET = 6,
 239   OP_UTF8_PERIOD = 7,
 240 #endif /* RE_ENABLE_I18N */
 241
 242   /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
 243      when the debugger shows values of this enum type.  */
 244 #define EPSILON_BIT 8
 245   OP_OPEN_SUBEXP = EPSILON_BIT | 0,
 246   OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
 247   OP_ALT = EPSILON_BIT | 2,
 248   OP_DUP_ASTERISK = EPSILON_BIT | 3,
 249   ANCHOR = EPSILON_BIT | 4,
 250
 251   /* Tree type, these are used only by tree. */
 252   CONCAT = 16,
 253   SUBEXP = 17,
 254
 255   /* Token type, these are used only by token.  */
 256   OP_DUP_PLUS = 18,
 257   OP_DUP_QUESTION,
 258   OP_OPEN_BRACKET,
 259   OP_CLOSE_BRACKET,
 260   OP_CHARSET_RANGE,
 261   OP_OPEN_DUP_NUM,
 262   OP_CLOSE_DUP_NUM,
 263   OP_NON_MATCH_LIST,
 264   OP_OPEN_COLL_ELEM,
 265   OP_CLOSE_COLL_ELEM,
 266   OP_OPEN_EQUIV_CLASS,
 267   OP_CLOSE_EQUIV_CLASS,
 268   OP_OPEN_CHAR_CLASS,
 269   OP_CLOSE_CHAR_CLASS,
 270   OP_WORD,
 271   OP_NOTWORD,
 272   OP_SPACE,
 273   OP_NOTSPACE,
 274   BACK_SLASH
 275
 276 } re_token_type_t;
 277
 278 #ifdef RE_ENABLE_I18N
 279 typedef struct
 280 {
 281   /* Multibyte characters.  */
 282   wchar_t *mbchars;
 283
 284   /* Collating symbols.  */
 285 # ifdef _LIBC
 286   int32_t *coll_syms;
 287 # endif
 288
 289   /* Equivalence classes. */
 290 # ifdef _LIBC
 291   int32_t *equiv_classes;
 292 # endif
 293
 294   /* Range expressions. */
 295 # ifdef _LIBC
 296   uint32_t *range_starts;
 297   uint32_t *range_ends;
 298 # else /* not _LIBC */
 299   wchar_t *range_starts;
 300   wchar_t *range_ends;
 301 # endif /* not _LIBC */
 302
 303   /* Character classes. */
 304   wctype_t *char_classes;
 305
 306   /* If this character set is the non-matching list.  */
 307   unsigned int non_match : 1;
 308
 309   /* # of multibyte characters.  */
 310   Idx nmbchars;
 311
 312   /* # of collating symbols.  */
 313   Idx ncoll_syms;
 314
 315   /* # of equivalence classes. */
 316   Idx nequiv_classes;
 317
 318   /* # of range expressions. */
 319   Idx nranges;
 320
 321   /* # of character classes. */
 322   Idx nchar_classes;
 323 } re_charset_t;
 324 #endif /* RE_ENABLE_I18N */
 325
 326 typedef struct
 327 {
 328   union
 329   {
 330     unsigned char c;            /* for CHARACTER */
 331     re_bitset_ptr_t sbcset;     /* for SIMPLE_BRACKET */
 332 #ifdef RE_ENABLE_I18N
 333     re_charset_t *mbcset;       /* for COMPLEX_BRACKET */
 334 #endif /* RE_ENABLE_I18N */
 335     Idx idx;                    /* for BACK_REF */
 336     re_context_type ctx_type;   /* for ANCHOR */
 337   } opr;
 338 #if __GNUC__ >= 2 && !defined __STRICT_ANSI__
 339   re_token_type_t type : 8;
 340 #else
 341   re_token_type_t type;
 342 #endif
 343   unsigned int constraint : 10; /* context constraint */
 344   unsigned int duplicated : 1;
 345   unsigned int opt_subexp : 1;
 346 #ifdef RE_ENABLE_I18N
 347   unsigned int accept_mb : 1;
 348   /* These 2 bits can be moved into the union if needed (e.g. if running out
 349      of bits; move opr.c to opr.c.c and move the flags to opr.c.flags).  */
 350   unsigned int mb_partial : 1;
 351 #endif
 352   unsigned int word_char : 1;
 353 } re_token_t;
 354
 355 #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
 356
 357 struct re_string_t
 358 {
 359   /* Indicate the raw buffer which is the original string passed as an
 360      argument of regexec(), re_search(), etc..  */
 361   const unsigned char *raw_mbs;
 362   /* Store the multibyte string.  In case of "case insensitive mode" like
 363      REG_ICASE, upper cases of the string are stored, otherwise MBS points
 364      the same address that RAW_MBS points.  */
 365   unsigned char *mbs;
 366 #ifdef RE_ENABLE_I18N
 367   /* Store the wide character string which is corresponding to MBS.  */
 368   wint_t *wcs;
 369   Idx *offsets;
 370   mbstate_t cur_state;
 371 #endif
 372   /* Index in RAW_MBS.  Each character mbs[i] corresponds to
 373      raw_mbs[raw_mbs_idx + i].  */
 374   Idx raw_mbs_idx;
 375   /* The length of the valid characters in the buffers.  */
 376   Idx valid_len;
 377   /* The corresponding number of bytes in raw_mbs array.  */
 378   Idx valid_raw_len;
 379   /* The length of the buffers MBS and WCS.  */
 380   Idx bufs_len;
 381   /* The index in MBS, which is updated by re_string_fetch_byte.  */
 382   Idx cur_idx;
 383   /* length of RAW_MBS array.  */
 384   Idx raw_len;
 385   /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN.  */
 386   Idx len;
 387   /* End of the buffer may be shorter than its length in the cases such
 388      as re_match_2, re_search_2.  Then, we use STOP for end of the buffer
 389      instead of LEN.  */
 390   Idx raw_stop;
 391   /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS.  */
 392   Idx stop;
 393
 394   /* The context of mbs[0].  We store the context independently, since
 395      the context of mbs[0] may be different from raw_mbs[0], which is
 396      the beginning of the input string.  */
 397   unsigned int tip_context;
 398   /* The translation passed as a part of an argument of re_compile_pattern.  */
 399   RE_TRANSLATE_TYPE trans;
 400   /* Copy of re_dfa_t's word_char.  */
 401   re_const_bitset_ptr_t word_char;
 402   /* true if REG_ICASE.  */
 403   unsigned char icase;
 404   unsigned char is_utf8;
 405   unsigned char map_notascii;
 406   unsigned char mbs_allocated;
 407   unsigned char offsets_needed;
 408   unsigned char newline_anchor;
 409   unsigned char word_ops_used;
 410   int mb_cur_max;
 411 };
 412 typedef struct re_string_t re_string_t;
 413
 414
 415 struct re_dfa_t;
 416 typedef struct re_dfa_t re_dfa_t;
 417
 418 #ifndef _LIBC
 419 # define IS_IN(libc) false
 420 #endif
 421
 422 #define re_string_peek_byte(pstr, offset) \
 423   ((pstr)->mbs[(pstr)->cur_idx + offset])
 424 #define re_string_fetch_byte(pstr) \
 425   ((pstr)->mbs[(pstr)->cur_idx++])
 426 #define re_string_first_byte(pstr, idx) \
 427   ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
 428 #define re_string_is_single_byte_char(pstr, idx) \
 429   ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
 430                                 || (pstr)->wcs[(idx) + 1] != WEOF))
 431 #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
 432 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
 433 #define re_string_get_buffer(pstr) ((pstr)->mbs)
 434 #define re_string_length(pstr) ((pstr)->len)
 435 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
 436 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
 437 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
 438
 439 #if defined _LIBC || HAVE_ALLOCA
 440 # include <alloca.h>
 441 #endif
 442
 443 #ifndef _LIBC
 444 # if HAVE_ALLOCA
 445 /* The OS usually guarantees only one guard page at the bottom of the stack,
 446    and a page size can be as small as 4096 bytes.  So we cannot safely
 447    allocate anything larger than 4096 bytes.  Also care for the possibility
 448    of a few compiler-allocated temporary stack slots.  */
 449 #  define __libc_use_alloca(n) ((n) < 4032)
 450 # else
 451 /* alloca is implemented with malloc, so just use malloc.  */
 452 #  define __libc_use_alloca(n) 0
 453 #  undef alloca
 454 #  define alloca(n) malloc (n)
 455 # endif
 456 #endif
 457
 458 #ifdef _LIBC
 459 # define MALLOC_0_IS_NONNULL 1
 460 #elif !defined MALLOC_0_IS_NONNULL
 461 # define MALLOC_0_IS_NONNULL 0
 462 #endif
 463
 464 #ifndef MAX
 465 # define MAX(a,b) ((a) < (b) ? (b) : (a))
 466 #endif
 467 #ifndef MIN
 468 # define MIN(a,b) ((a) < (b) ? (a) : (b))
 469 #endif
 470
 471 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
 472 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
 473 #define re_free(p) free (p)
 474
 475 struct bin_tree_t
 476 {
 477   struct bin_tree_t *parent;
 478   struct bin_tree_t *left;
 479   struct bin_tree_t *right;
 480   struct bin_tree_t *first;
 481   struct bin_tree_t *next;
 482
 483   re_token_t token;
 484
 485   /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
 486      Otherwise 'type' indicate the type of this node.  */
 487   Idx node_idx;
 488 };
 489 typedef struct bin_tree_t bin_tree_t;
 490
 491 #define BIN_TREE_STORAGE_SIZE \
 492   ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
 493
 494 struct bin_tree_storage_t
 495 {
 496   struct bin_tree_storage_t *next;
 497   bin_tree_t data[BIN_TREE_STORAGE_SIZE];
 498 };
 499 typedef struct bin_tree_storage_t bin_tree_storage_t;
 500
 501 #define CONTEXT_WORD 1
 502 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
 503 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
 504 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
 505
 506 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
 507 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
 508 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
 509 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
 510 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
 511
 512 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
 513 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
 514 #define IS_WIDE_WORD_CHAR(ch) (__iswalnum (ch) || (ch) == L'_')
 515 #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
 516
 517 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
 518  ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
 519   || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
 520   || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
 521   || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
 522
 523 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
 524  ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
 525   || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
 526   || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
 527   || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
 528
 529 struct re_dfastate_t
 530 {
 531   re_hashval_t hash;
 532   re_node_set nodes;
 533   re_node_set non_eps_nodes;
 534   re_node_set inveclosure;
 535   re_node_set *entrance_nodes;
 536   struct re_dfastate_t **trtable, **word_trtable;
 537   unsigned int context : 4;
 538   unsigned int halt : 1;
 539   /* If this state can accept "multi byte".
 540      Note that we refer to multibyte characters, and multi character
 541      collating elements as "multi byte".  */
 542   unsigned int accept_mb : 1;
 543   /* If this state has backreference node(s).  */
 544   unsigned int has_backref : 1;
 545   unsigned int has_constraint : 1;
 546 };
 547 typedef struct re_dfastate_t re_dfastate_t;
 548
 549 struct re_state_table_entry
 550 {
 551   Idx num;
 552   Idx alloc;
 553   re_dfastate_t **array;
 554 };
 555
 556 /* Array type used in re_sub_match_last_t and re_sub_match_top_t.  */
 557
 558 typedef struct
 559 {
 560   Idx next_idx;
 561   Idx alloc;
 562   re_dfastate_t **array;
 563 } state_array_t;
 564
 565 /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP.  */
 566
 567 typedef struct
 568 {
 569   Idx node;
 570   Idx str_idx; /* The position NODE match at.  */
 571   state_array_t path;
 572 } re_sub_match_last_t;
 573
 574 /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
 575    And information about the node, whose type is OP_CLOSE_SUBEXP,
 576    corresponding to NODE is stored in LASTS.  */
 577
 578 typedef struct
 579 {
 580   Idx str_idx;
 581   Idx node;
 582   state_array_t *path;
 583   Idx alasts; /* Allocation size of LASTS.  */
 584   Idx nlasts; /* The number of LASTS.  */
 585   re_sub_match_last_t **lasts;
 586 } re_sub_match_top_t;
 587
 588 struct re_backref_cache_entry
 589 {
 590   Idx node;
 591   Idx str_idx;
 592   Idx subexp_from;
 593   Idx subexp_to;
 594   bitset_word_t eps_reachable_subexps_map;
 595   char more;
 596 };
 597
 598 typedef struct
 599 {
 600   /* The string object corresponding to the input string.  */
 601   re_string_t input;
 602   const re_dfa_t *const dfa;
 603   /* EFLAGS of the argument of regexec.  */
 604   int eflags;
 605   /* Where the matching ends.  */
 606   Idx match_last;
 607   Idx last_node;
 608   /* The state log used by the matcher.  */
 609   re_dfastate_t **state_log;
 610   Idx state_log_top;
 611   /* Back reference cache.  */
 612   Idx nbkref_ents;
 613   Idx abkref_ents;
 614   struct re_backref_cache_entry *bkref_ents;
 615   int max_mb_elem_len;
 616   Idx nsub_tops;
 617   Idx asub_tops;
 618   re_sub_match_top_t **sub_tops;
 619 } re_match_context_t;
 620
 621 typedef struct
 622 {
 623   re_dfastate_t **sifted_states;
 624   re_dfastate_t **limited_states;
 625   Idx last_node;
 626   Idx last_str_idx;
 627   re_node_set limits;
 628 } re_sift_context_t;
 629
 630 struct re_fail_stack_ent_t
 631 {
 632   Idx idx;
 633   Idx node;
 634   regmatch_t *regs;
 635   re_node_set eps_via_nodes;
 636 };
 637
 638 struct re_fail_stack_t
 639 {
 640   Idx num;
 641   Idx alloc;
 642   struct re_fail_stack_ent_t *stack;
 643 };
 644
 645 struct re_dfa_t
 646 {
 647   re_token_t *nodes;
 648   size_t nodes_alloc;
 649   size_t nodes_len;
 650   Idx *nexts;
 651   Idx *org_indices;
 652   re_node_set *edests;
 653   re_node_set *eclosures;
 654   re_node_set *inveclosures;
 655   struct re_state_table_entry *state_table;
 656   re_dfastate_t *init_state;
 657   re_dfastate_t *init_state_word;
 658   re_dfastate_t *init_state_nl;
 659   re_dfastate_t *init_state_begbuf;
 660   bin_tree_t *str_tree;
 661   bin_tree_storage_t *str_tree_storage;
 662   re_bitset_ptr_t sb_char;
 663   int str_tree_storage_idx;
 664
 665   /* number of subexpressions 're_nsub' is in regex_t.  */
 666   re_hashval_t state_hash_mask;
 667   Idx init_node;
 668   Idx nbackref; /* The number of backreference in this dfa.  */
 669
 670   /* Bitmap expressing which backreference is used.  */
 671   bitset_word_t used_bkref_map;
 672   bitset_word_t completed_bkref_map;
 673
 674   unsigned int has_plural_match : 1;
 675   /* If this dfa has "multibyte node", which is a backreference or
 676      a node which can accept multibyte character or multi character
 677      collating element.  */
 678   unsigned int has_mb_node : 1;
 679   unsigned int is_utf8 : 1;
 680   unsigned int map_notascii : 1;
 681   unsigned int word_ops_used : 1;
 682   int mb_cur_max;
 683   bitset_t word_char;
 684   reg_syntax_t syntax;
 685   Idx *subexp_map;
 686 #ifdef DEBUG
 687   char* re_str;
 688 #endif
 689   lock_define (lock)
 690 };
 691
 692 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
 693 #define re_node_set_remove(set,id) \
 694   (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
 695 #define re_node_set_empty(p) ((p)->nelem = 0)
 696 #define re_node_set_free(set) re_free ((set)->elems)
 697 \f
 698
 699 typedef enum
 700 {
 701   SB_CHAR,
 702   MB_CHAR,
 703   EQUIV_CLASS,
 704   COLL_SYM,
 705   CHAR_CLASS
 706 } bracket_elem_type;
 707
 708 typedef struct
 709 {
 710   bracket_elem_type type;
 711   union
 712   {
 713     unsigned char ch;
 714     unsigned char *name;
 715     wchar_t wch;
 716   } opr;
 717 } bracket_elem_t;
 718
 719
 720 /* Functions for bitset_t operation.  */
 721
 722 static inline void
 723 bitset_set (bitset_t set, Idx i)
 724 {
 725   set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
 726 }
 727
 728 static inline void
 729 bitset_clear (bitset_t set, Idx i)
 730 {
 731   set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
 732 }
 733
 734 static inline bool
 735 bitset_contain (const bitset_t set, Idx i)
 736 {
 737   return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
 738 }
 739
 740 static inline void
 741 bitset_empty (bitset_t set)
 742 {
 743   memset (set, '\0', sizeof (bitset_t));
 744 }
 745
 746 static inline void
 747 bitset_set_all (bitset_t set)
 748 {
 749   memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
 750   if (SBC_MAX % BITSET_WORD_BITS != 0)
 751     set[BITSET_WORDS - 1] =
 752       ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
 753 }
 754
 755 static inline void
 756 bitset_copy (bitset_t dest, const bitset_t src)
 757 {
 758   memcpy (dest, src, sizeof (bitset_t));
 759 }
 760
 761 static inline void
 762 bitset_not (bitset_t set)
 763 {
 764   int bitset_i;
 765   for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
 766     set[bitset_i] = ~set[bitset_i];
 767   if (SBC_MAX % BITSET_WORD_BITS != 0)
 768     set[BITSET_WORDS - 1] =
 769       ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
 770        & ~set[BITSET_WORDS - 1]);
 771 }
 772
 773 static inline void
 774 bitset_merge (bitset_t dest, const bitset_t src)
 775 {
 776   int bitset_i;
 777   for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
 778     dest[bitset_i] |= src[bitset_i];
 779 }
 780
 781 static inline void
 782 bitset_mask (bitset_t dest, const bitset_t src)
 783 {
 784   int bitset_i;
 785   for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
 786     dest[bitset_i] &= src[bitset_i];
 787 }
 788
 789 #ifdef RE_ENABLE_I18N
 790 /* Functions for re_string.  */
 791 static int
 792 __attribute__ ((pure, unused))
 793 re_string_char_size_at (const re_string_t *pstr, Idx idx)
 794 {
 795   int byte_idx;
 796   if (pstr->mb_cur_max == 1)
 797     return 1;
 798   for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
 799     if (pstr->wcs[idx + byte_idx] != WEOF)
 800       break;
 801   return byte_idx;
 802 }
 803
 804 static wint_t
 805 __attribute__ ((pure, unused))
 806 re_string_wchar_at (const re_string_t *pstr, Idx idx)
 807 {
 808   if (pstr->mb_cur_max == 1)
 809     return (wint_t) pstr->mbs[idx];
 810   return (wint_t) pstr->wcs[idx];
 811 }
 812
 813 # ifdef _LIBC
 814 #  include <locale/weight.h>
 815 # endif
 816
 817 static int
 818 __attribute__ ((pure, unused))
 819 re_string_elem_size_at (const re_string_t *pstr, Idx idx)
 820 {
 821 # ifdef _LIBC
 822   const unsigned char *p, *extra;
 823   const int32_t *table, *indirect;
 824   uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
 825
 826   if (nrules != 0)
 827     {
 828       table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
 829       extra = (const unsigned char *)
 830         _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
 831       indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
 832                                                 _NL_COLLATE_INDIRECTMB);
 833       p = pstr->mbs + idx;
 834       findidx (table, indirect, extra, &p, pstr->len - idx);
 835       return p - pstr->mbs - idx;
 836     }
 837   else
 838 # endif /* _LIBC */
 839     return 1;
 840 }
 841 #endif /* RE_ENABLE_I18N */
 842
 843 #ifndef FALLTHROUGH
 844 # if __GNUC__ < 7
 845 #  define FALLTHROUGH ((void) 0)
 846 # else
 847 #  define FALLTHROUGH __attribute__ ((__fallthrough__))
 848 # endif
 849 #endif
 850
 851 #endif /*  _REGEX_INTERNAL_H */