| 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 | #if !defined(KMK_GREP) || !defined(_MSC_VER)
|
|---|
| 29 | #include <langinfo.h>
|
|---|
| 30 | #endif
|
|---|
| 31 | #include <locale.h>
|
|---|
| 32 | #include <wchar.h>
|
|---|
| 33 | #include <wctype.h>
|
|---|
| 34 | #include <stdbool.h>
|
|---|
| 35 | #include <stdint.h>
|
|---|
| 36 |
|
|---|
| 37 | #ifndef _LIBC
|
|---|
| 38 | # include <dynarray.h>
|
|---|
| 39 | #endif
|
|---|
| 40 |
|
|---|
| 41 | #include <intprops.h>
|
|---|
| 42 | #include <verify.h>
|
|---|
| 43 |
|
|---|
| 44 | #if defined DEBUG && DEBUG != 0
|
|---|
| 45 | # include <assert.h>
|
|---|
| 46 | # define DEBUG_ASSERT(x) assert (x)
|
|---|
| 47 | #else
|
|---|
| 48 | # define DEBUG_ASSERT(x) assume (x)
|
|---|
| 49 | #endif
|
|---|
| 50 |
|
|---|
| 51 | #ifdef _LIBC
|
|---|
| 52 | # include <libc-lock.h>
|
|---|
| 53 | # define lock_define(name) __libc_lock_define (, name)
|
|---|
| 54 | # define lock_init(lock) (__libc_lock_init (lock), 0)
|
|---|
| 55 | # define lock_fini(lock) ((void) 0)
|
|---|
| 56 | # define lock_lock(lock) __libc_lock_lock (lock)
|
|---|
| 57 | # define lock_unlock(lock) __libc_lock_unlock (lock)
|
|---|
| 58 | #elif defined GNULIB_LOCK && !defined GNULIB_REGEX_SINGLE_THREAD
|
|---|
| 59 | # include "glthread/lock.h"
|
|---|
| 60 | # define lock_define(name) gl_lock_define (, name)
|
|---|
| 61 | # define lock_init(lock) glthread_lock_init (&(lock))
|
|---|
| 62 | # define lock_fini(lock) glthread_lock_destroy (&(lock))
|
|---|
| 63 | # define lock_lock(lock) glthread_lock_lock (&(lock))
|
|---|
| 64 | # define lock_unlock(lock) glthread_lock_unlock (&(lock))
|
|---|
| 65 | #elif defined GNULIB_PTHREAD && !defined GNULIB_REGEX_SINGLE_THREAD
|
|---|
| 66 | # include <pthread.h>
|
|---|
| 67 | # define lock_define(name) pthread_mutex_t name;
|
|---|
| 68 | # define lock_init(lock) pthread_mutex_init (&(lock), 0)
|
|---|
| 69 | # define lock_fini(lock) pthread_mutex_destroy (&(lock))
|
|---|
| 70 | # define lock_lock(lock) pthread_mutex_lock (&(lock))
|
|---|
| 71 | # define lock_unlock(lock) pthread_mutex_unlock (&(lock))
|
|---|
| 72 | #else
|
|---|
| 73 | # define lock_define(name)
|
|---|
| 74 | # define lock_init(lock) 0
|
|---|
| 75 | # define lock_fini(lock) ((void) 0)
|
|---|
| 76 | /* The 'dfa' avoids an "unused variable 'dfa'" warning from GCC. */
|
|---|
| 77 | # define lock_lock(lock) ((void) dfa)
|
|---|
| 78 | # define lock_unlock(lock) ((void) 0)
|
|---|
| 79 | #endif
|
|---|
| 80 |
|
|---|
| 81 | /* In case that the system doesn't have isblank(). */
|
|---|
| 82 | #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
|
|---|
| 83 | # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
|
|---|
| 84 | #endif
|
|---|
| 85 |
|
|---|
| 86 | /* regex code assumes isascii has its usual numeric meaning,
|
|---|
| 87 | even if the portable character set uses EBCDIC encoding,
|
|---|
| 88 | and even if wint_t is wider than int. */
|
|---|
| 89 | #ifndef _LIBC
|
|---|
| 90 | # undef isascii
|
|---|
| 91 | # define isascii(c) (((c) & ~0x7f) == 0)
|
|---|
| 92 | #endif
|
|---|
| 93 |
|
|---|
| 94 | #ifdef _LIBC
|
|---|
| 95 | # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
|
|---|
| 96 | # define _RE_DEFINE_LOCALE_FUNCTIONS 1
|
|---|
| 97 | # include <locale/localeinfo.h>
|
|---|
| 98 | # include <locale/coll-lookup.h>
|
|---|
| 99 | # endif
|
|---|
| 100 | #endif
|
|---|
| 101 |
|
|---|
| 102 | /* This is for other GNU distributions with internationalized messages. */
|
|---|
| 103 | #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
|
|---|
| 104 | # include <libintl.h>
|
|---|
| 105 | # ifdef _LIBC
|
|---|
| 106 | # undef gettext
|
|---|
| 107 | # define gettext(msgid) \
|
|---|
| 108 | __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES)
|
|---|
| 109 | # endif
|
|---|
| 110 | #else
|
|---|
| 111 | # undef gettext
|
|---|
| 112 | # define gettext(msgid) (msgid)
|
|---|
| 113 | #endif
|
|---|
| 114 |
|
|---|
| 115 | #ifndef gettext_noop
|
|---|
| 116 | /* This define is so xgettext can find the internationalizable
|
|---|
| 117 | strings. */
|
|---|
| 118 | # define gettext_noop(String) String
|
|---|
| 119 | #endif
|
|---|
| 120 |
|
|---|
| 121 | #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE) || _LIBC
|
|---|
| 122 | # define RE_ENABLE_I18N
|
|---|
| 123 | #endif
|
|---|
| 124 |
|
|---|
| 125 | /* Number of ASCII characters. */
|
|---|
| 126 | #define ASCII_CHARS 0x80
|
|---|
| 127 |
|
|---|
| 128 | /* Number of single byte characters. */
|
|---|
| 129 | #define SBC_MAX (UCHAR_MAX + 1)
|
|---|
| 130 |
|
|---|
| 131 | #define COLL_ELEM_LEN_MAX 8
|
|---|
| 132 |
|
|---|
| 133 | /* The character which represents newline. */
|
|---|
| 134 | #define NEWLINE_CHAR '\n'
|
|---|
| 135 | #define WIDE_NEWLINE_CHAR L'\n'
|
|---|
| 136 |
|
|---|
| 137 | /* Rename to standard API for using out of glibc. */
|
|---|
| 138 | #ifndef _LIBC
|
|---|
| 139 | # undef __wctype
|
|---|
| 140 | # undef __iswalnum
|
|---|
| 141 | # undef __iswctype
|
|---|
| 142 | # undef __towlower
|
|---|
| 143 | # undef __towupper
|
|---|
| 144 | # define __wctype wctype
|
|---|
| 145 | # define __iswalnum iswalnum
|
|---|
| 146 | # define __iswctype iswctype
|
|---|
| 147 | # define __towlower towlower
|
|---|
| 148 | # define __towupper towupper
|
|---|
| 149 | # define __btowc btowc
|
|---|
| 150 | # define __mbrtowc mbrtowc
|
|---|
| 151 | # define __wcrtomb wcrtomb
|
|---|
| 152 | # define __regfree regfree
|
|---|
| 153 | #endif /* not _LIBC */
|
|---|
| 154 |
|
|---|
| 155 | #ifndef SSIZE_MAX
|
|---|
| 156 | # define SSIZE_MAX ((ssize_t) (SIZE_MAX / 2))
|
|---|
| 157 | #endif
|
|---|
| 158 | #ifndef ULONG_WIDTH
|
|---|
| 159 | # define ULONG_WIDTH REGEX_UINTEGER_WIDTH (ULONG_MAX)
|
|---|
| 160 | /* The number of usable bits in an unsigned integer type with maximum
|
|---|
| 161 | value MAX, as an int expression suitable in #if. Cover all known
|
|---|
| 162 | practical hosts. This implementation exploits the fact that MAX is
|
|---|
| 163 | 1 less than a power of 2, and merely counts the number of 1 bits in
|
|---|
| 164 | MAX; "COBn" means "count the number of 1 bits in the low-order n bits". */
|
|---|
| 165 | # define REGEX_UINTEGER_WIDTH(max) REGEX_COB128 (max)
|
|---|
| 166 | # define REGEX_COB128(n) (REGEX_COB64 ((n) >> 31 >> 31 >> 2) + REGEX_COB64 (n))
|
|---|
| 167 | # define REGEX_COB64(n) (REGEX_COB32 ((n) >> 31 >> 1) + REGEX_COB32 (n))
|
|---|
| 168 | # define REGEX_COB32(n) (REGEX_COB16 ((n) >> 16) + REGEX_COB16 (n))
|
|---|
| 169 | # define REGEX_COB16(n) (REGEX_COB8 ((n) >> 8) + REGEX_COB8 (n))
|
|---|
| 170 | # define REGEX_COB8(n) (REGEX_COB4 ((n) >> 4) + REGEX_COB4 (n))
|
|---|
| 171 | # define REGEX_COB4(n) (!!((n) & 8) + !!((n) & 4) + !!((n) & 2) + ((n) & 1))
|
|---|
| 172 | # if ULONG_MAX / 2 + 1 != 1ul << (ULONG_WIDTH - 1)
|
|---|
| 173 | # error "ULONG_MAX out of range"
|
|---|
| 174 | # endif
|
|---|
| 175 | #endif
|
|---|
| 176 |
|
|---|
| 177 | /* The type of indexes into strings. This is signed, not size_t,
|
|---|
| 178 | since the API requires indexes to fit in regoff_t anyway, and using
|
|---|
| 179 | signed integers makes the code a bit smaller and presumably faster.
|
|---|
| 180 | The traditional GNU regex implementation uses int for indexes.
|
|---|
| 181 | The POSIX-compatible implementation uses a possibly-wider type.
|
|---|
| 182 | The name 'Idx' is three letters to minimize the hassle of
|
|---|
| 183 | reindenting a lot of regex code that formerly used 'int'. */
|
|---|
| 184 | typedef regoff_t Idx;
|
|---|
| 185 | #ifdef _REGEX_LARGE_OFFSETS
|
|---|
| 186 | # define IDX_MAX SSIZE_MAX
|
|---|
| 187 | #else
|
|---|
| 188 | # define IDX_MAX INT_MAX
|
|---|
| 189 | #endif
|
|---|
| 190 |
|
|---|
| 191 | /* A hash value, suitable for computing hash tables. */
|
|---|
| 192 | typedef __re_size_t re_hashval_t;
|
|---|
| 193 |
|
|---|
| 194 | /* An integer used to represent a set of bits. It must be unsigned,
|
|---|
| 195 | and must be at least as wide as unsigned int. */
|
|---|
| 196 | typedef unsigned long int bitset_word_t;
|
|---|
| 197 | /* All bits set in a bitset_word_t. */
|
|---|
| 198 | #define BITSET_WORD_MAX ULONG_MAX
|
|---|
| 199 | /* Number of bits in a bitset_word_t. */
|
|---|
| 200 | #define BITSET_WORD_BITS ULONG_WIDTH
|
|---|
| 201 |
|
|---|
| 202 | /* Number of bitset_word_t values in a bitset_t. */
|
|---|
| 203 | #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
|
|---|
| 204 |
|
|---|
| 205 | typedef bitset_word_t bitset_t[BITSET_WORDS];
|
|---|
| 206 | typedef bitset_word_t *re_bitset_ptr_t;
|
|---|
| 207 | typedef const bitset_word_t *re_const_bitset_ptr_t;
|
|---|
| 208 |
|
|---|
| 209 | #define PREV_WORD_CONSTRAINT 0x0001
|
|---|
| 210 | #define PREV_NOTWORD_CONSTRAINT 0x0002
|
|---|
| 211 | #define NEXT_WORD_CONSTRAINT 0x0004
|
|---|
| 212 | #define NEXT_NOTWORD_CONSTRAINT 0x0008
|
|---|
| 213 | #define PREV_NEWLINE_CONSTRAINT 0x0010
|
|---|
| 214 | #define NEXT_NEWLINE_CONSTRAINT 0x0020
|
|---|
| 215 | #define PREV_BEGBUF_CONSTRAINT 0x0040
|
|---|
| 216 | #define NEXT_ENDBUF_CONSTRAINT 0x0080
|
|---|
| 217 | #define WORD_DELIM_CONSTRAINT 0x0100
|
|---|
| 218 | #define NOT_WORD_DELIM_CONSTRAINT 0x0200
|
|---|
| 219 |
|
|---|
| 220 | typedef enum
|
|---|
| 221 | {
|
|---|
| 222 | INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
|
|---|
| 223 | WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
|
|---|
| 224 | WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
|
|---|
| 225 | INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
|
|---|
| 226 | LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
|
|---|
| 227 | LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
|
|---|
| 228 | BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
|
|---|
| 229 | BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
|
|---|
| 230 | WORD_DELIM = WORD_DELIM_CONSTRAINT,
|
|---|
| 231 | NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
|
|---|
| 232 | } re_context_type;
|
|---|
| 233 |
|
|---|
| 234 | typedef struct
|
|---|
| 235 | {
|
|---|
| 236 | Idx alloc;
|
|---|
| 237 | Idx nelem;
|
|---|
| 238 | Idx *elems;
|
|---|
| 239 | } re_node_set;
|
|---|
| 240 |
|
|---|
| 241 | typedef enum
|
|---|
| 242 | {
|
|---|
| 243 | NON_TYPE = 0,
|
|---|
| 244 |
|
|---|
| 245 | /* Node type, These are used by token, node, tree. */
|
|---|
| 246 | CHARACTER = 1,
|
|---|
| 247 | END_OF_RE = 2,
|
|---|
| 248 | SIMPLE_BRACKET = 3,
|
|---|
| 249 | OP_BACK_REF = 4,
|
|---|
| 250 | OP_PERIOD = 5,
|
|---|
| 251 | #ifdef RE_ENABLE_I18N
|
|---|
| 252 | COMPLEX_BRACKET = 6,
|
|---|
| 253 | OP_UTF8_PERIOD = 7,
|
|---|
| 254 | #endif /* RE_ENABLE_I18N */
|
|---|
| 255 |
|
|---|
| 256 | /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
|
|---|
| 257 | when the debugger shows values of this enum type. */
|
|---|
| 258 | #define EPSILON_BIT 8
|
|---|
| 259 | OP_OPEN_SUBEXP = EPSILON_BIT | 0,
|
|---|
| 260 | OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
|
|---|
| 261 | OP_ALT = EPSILON_BIT | 2,
|
|---|
| 262 | OP_DUP_ASTERISK = EPSILON_BIT | 3,
|
|---|
| 263 | ANCHOR = EPSILON_BIT | 4,
|
|---|
| 264 |
|
|---|
| 265 | /* Tree type, these are used only by tree. */
|
|---|
| 266 | CONCAT = 16,
|
|---|
| 267 | SUBEXP = 17,
|
|---|
| 268 |
|
|---|
| 269 | /* Token type, these are used only by token. */
|
|---|
| 270 | OP_DUP_PLUS = 18,
|
|---|
| 271 | OP_DUP_QUESTION,
|
|---|
| 272 | OP_OPEN_BRACKET,
|
|---|
| 273 | OP_CLOSE_BRACKET,
|
|---|
| 274 | OP_CHARSET_RANGE,
|
|---|
| 275 | OP_OPEN_DUP_NUM,
|
|---|
| 276 | OP_CLOSE_DUP_NUM,
|
|---|
| 277 | OP_NON_MATCH_LIST,
|
|---|
| 278 | OP_OPEN_COLL_ELEM,
|
|---|
| 279 | OP_CLOSE_COLL_ELEM,
|
|---|
| 280 | OP_OPEN_EQUIV_CLASS,
|
|---|
| 281 | OP_CLOSE_EQUIV_CLASS,
|
|---|
| 282 | OP_OPEN_CHAR_CLASS,
|
|---|
| 283 | OP_CLOSE_CHAR_CLASS,
|
|---|
| 284 | OP_WORD,
|
|---|
| 285 | OP_NOTWORD,
|
|---|
| 286 | OP_SPACE,
|
|---|
| 287 | OP_NOTSPACE,
|
|---|
| 288 | BACK_SLASH
|
|---|
| 289 |
|
|---|
| 290 | } re_token_type_t;
|
|---|
| 291 |
|
|---|
| 292 | #ifdef RE_ENABLE_I18N
|
|---|
| 293 | typedef struct
|
|---|
| 294 | {
|
|---|
| 295 | /* Multibyte characters. */
|
|---|
| 296 | wchar_t *mbchars;
|
|---|
| 297 |
|
|---|
| 298 | /* Collating symbols. */
|
|---|
| 299 | # ifdef _LIBC
|
|---|
| 300 | int32_t *coll_syms;
|
|---|
| 301 | # endif
|
|---|
| 302 |
|
|---|
| 303 | /* Equivalence classes. */
|
|---|
| 304 | # ifdef _LIBC
|
|---|
| 305 | int32_t *equiv_classes;
|
|---|
| 306 | # endif
|
|---|
| 307 |
|
|---|
| 308 | /* Range expressions. */
|
|---|
| 309 | # ifdef _LIBC
|
|---|
| 310 | uint32_t *range_starts;
|
|---|
| 311 | uint32_t *range_ends;
|
|---|
| 312 | # else /* not _LIBC */
|
|---|
| 313 | wchar_t *range_starts;
|
|---|
| 314 | wchar_t *range_ends;
|
|---|
| 315 | # endif /* not _LIBC */
|
|---|
| 316 |
|
|---|
| 317 | /* Character classes. */
|
|---|
| 318 | wctype_t *char_classes;
|
|---|
| 319 |
|
|---|
| 320 | /* If this character set is the non-matching list. */
|
|---|
| 321 | unsigned int non_match : 1;
|
|---|
| 322 |
|
|---|
| 323 | /* # of multibyte characters. */
|
|---|
| 324 | Idx nmbchars;
|
|---|
| 325 |
|
|---|
| 326 | /* # of collating symbols. */
|
|---|
| 327 | Idx ncoll_syms;
|
|---|
| 328 |
|
|---|
| 329 | /* # of equivalence classes. */
|
|---|
| 330 | Idx nequiv_classes;
|
|---|
| 331 |
|
|---|
| 332 | /* # of range expressions. */
|
|---|
| 333 | Idx nranges;
|
|---|
| 334 |
|
|---|
| 335 | /* # of character classes. */
|
|---|
| 336 | Idx nchar_classes;
|
|---|
| 337 | } re_charset_t;
|
|---|
| 338 | #endif /* RE_ENABLE_I18N */
|
|---|
| 339 |
|
|---|
| 340 | typedef struct
|
|---|
| 341 | {
|
|---|
| 342 | union
|
|---|
| 343 | {
|
|---|
| 344 | unsigned char c; /* for CHARACTER */
|
|---|
| 345 | re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
|
|---|
| 346 | #ifdef RE_ENABLE_I18N
|
|---|
| 347 | re_charset_t *mbcset; /* for COMPLEX_BRACKET */
|
|---|
| 348 | #endif /* RE_ENABLE_I18N */
|
|---|
| 349 | Idx idx; /* for BACK_REF */
|
|---|
| 350 | re_context_type ctx_type; /* for ANCHOR */
|
|---|
| 351 | } opr;
|
|---|
| 352 | #if (__GNUC__ >= 2 || defined __clang__) && !defined __STRICT_ANSI__
|
|---|
| 353 | re_token_type_t type : 8;
|
|---|
| 354 | #else
|
|---|
| 355 | re_token_type_t type;
|
|---|
| 356 | #endif
|
|---|
| 357 | unsigned int constraint : 10; /* context constraint */
|
|---|
| 358 | unsigned int duplicated : 1;
|
|---|
| 359 | unsigned int opt_subexp : 1;
|
|---|
| 360 | #ifdef RE_ENABLE_I18N
|
|---|
| 361 | unsigned int accept_mb : 1;
|
|---|
| 362 | /* These 2 bits can be moved into the union if needed (e.g. if running out
|
|---|
| 363 | of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
|
|---|
| 364 | unsigned int mb_partial : 1;
|
|---|
| 365 | #endif
|
|---|
| 366 | unsigned int word_char : 1;
|
|---|
| 367 | } re_token_t;
|
|---|
| 368 |
|
|---|
| 369 | #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
|
|---|
| 370 |
|
|---|
| 371 | struct re_string_t
|
|---|
| 372 | {
|
|---|
| 373 | /* Indicate the raw buffer which is the original string passed as an
|
|---|
| 374 | argument of regexec(), re_search(), etc.. */
|
|---|
| 375 | const unsigned char *raw_mbs;
|
|---|
| 376 | /* Store the multibyte string. In case of "case insensitive mode" like
|
|---|
| 377 | REG_ICASE, upper cases of the string are stored, otherwise MBS points
|
|---|
| 378 | the same address that RAW_MBS points. */
|
|---|
| 379 | unsigned char *mbs;
|
|---|
| 380 | #ifdef RE_ENABLE_I18N
|
|---|
| 381 | /* Store the wide character string which is corresponding to MBS. */
|
|---|
| 382 | wint_t *wcs;
|
|---|
| 383 | Idx *offsets;
|
|---|
| 384 | mbstate_t cur_state;
|
|---|
| 385 | #endif
|
|---|
| 386 | /* Index in RAW_MBS. Each character mbs[i] corresponds to
|
|---|
| 387 | raw_mbs[raw_mbs_idx + i]. */
|
|---|
| 388 | Idx raw_mbs_idx;
|
|---|
| 389 | /* The length of the valid characters in the buffers. */
|
|---|
| 390 | Idx valid_len;
|
|---|
| 391 | /* The corresponding number of bytes in raw_mbs array. */
|
|---|
| 392 | Idx valid_raw_len;
|
|---|
| 393 | /* The length of the buffers MBS and WCS. */
|
|---|
| 394 | Idx bufs_len;
|
|---|
| 395 | /* The index in MBS, which is updated by re_string_fetch_byte. */
|
|---|
| 396 | Idx cur_idx;
|
|---|
| 397 | /* length of RAW_MBS array. */
|
|---|
| 398 | Idx raw_len;
|
|---|
| 399 | /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
|
|---|
| 400 | Idx len;
|
|---|
| 401 | /* End of the buffer may be shorter than its length in the cases such
|
|---|
| 402 | as re_match_2, re_search_2. Then, we use STOP for end of the buffer
|
|---|
| 403 | instead of LEN. */
|
|---|
| 404 | Idx raw_stop;
|
|---|
| 405 | /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
|
|---|
| 406 | Idx stop;
|
|---|
| 407 |
|
|---|
| 408 | /* The context of mbs[0]. We store the context independently, since
|
|---|
| 409 | the context of mbs[0] may be different from raw_mbs[0], which is
|
|---|
| 410 | the beginning of the input string. */
|
|---|
| 411 | unsigned int tip_context;
|
|---|
| 412 | /* The translation passed as a part of an argument of re_compile_pattern. */
|
|---|
| 413 | RE_TRANSLATE_TYPE trans;
|
|---|
| 414 | /* Copy of re_dfa_t's word_char. */
|
|---|
| 415 | re_const_bitset_ptr_t word_char;
|
|---|
| 416 | /* true if REG_ICASE. */
|
|---|
| 417 | unsigned char icase;
|
|---|
| 418 | unsigned char is_utf8;
|
|---|
| 419 | unsigned char map_notascii;
|
|---|
| 420 | unsigned char mbs_allocated;
|
|---|
| 421 | unsigned char offsets_needed;
|
|---|
| 422 | unsigned char newline_anchor;
|
|---|
| 423 | unsigned char word_ops_used;
|
|---|
| 424 | int mb_cur_max;
|
|---|
| 425 | };
|
|---|
| 426 | typedef struct re_string_t re_string_t;
|
|---|
| 427 |
|
|---|
| 428 |
|
|---|
| 429 | struct re_dfa_t;
|
|---|
| 430 | typedef struct re_dfa_t re_dfa_t;
|
|---|
| 431 |
|
|---|
| 432 | #ifndef _LIBC
|
|---|
| 433 | # define IS_IN(libc) false
|
|---|
| 434 | #endif
|
|---|
| 435 |
|
|---|
| 436 | #define re_string_peek_byte(pstr, offset) \
|
|---|
| 437 | ((pstr)->mbs[(pstr)->cur_idx + offset])
|
|---|
| 438 | #define re_string_fetch_byte(pstr) \
|
|---|
| 439 | ((pstr)->mbs[(pstr)->cur_idx++])
|
|---|
| 440 | #define re_string_first_byte(pstr, idx) \
|
|---|
| 441 | ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
|
|---|
| 442 | #define re_string_is_single_byte_char(pstr, idx) \
|
|---|
| 443 | ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
|
|---|
| 444 | || (pstr)->wcs[(idx) + 1] != WEOF))
|
|---|
| 445 | #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
|
|---|
| 446 | #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
|
|---|
| 447 | #define re_string_get_buffer(pstr) ((pstr)->mbs)
|
|---|
| 448 | #define re_string_length(pstr) ((pstr)->len)
|
|---|
| 449 | #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
|
|---|
| 450 | #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
|
|---|
| 451 | #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
|
|---|
| 452 |
|
|---|
| 453 | #ifdef _LIBC
|
|---|
| 454 | # define MALLOC_0_IS_NONNULL 1
|
|---|
| 455 | #elif !defined MALLOC_0_IS_NONNULL
|
|---|
| 456 | # define MALLOC_0_IS_NONNULL 0
|
|---|
| 457 | #endif
|
|---|
| 458 |
|
|---|
| 459 | #ifndef MAX
|
|---|
| 460 | # define MAX(a,b) ((a) < (b) ? (b) : (a))
|
|---|
| 461 | #endif
|
|---|
| 462 | #ifndef MIN
|
|---|
| 463 | # define MIN(a,b) ((a) < (b) ? (a) : (b))
|
|---|
| 464 | #endif
|
|---|
| 465 |
|
|---|
| 466 | #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
|
|---|
| 467 | #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
|
|---|
| 468 | #define re_free(p) free (p)
|
|---|
| 469 |
|
|---|
| 470 | struct bin_tree_t
|
|---|
| 471 | {
|
|---|
| 472 | struct bin_tree_t *parent;
|
|---|
| 473 | struct bin_tree_t *left;
|
|---|
| 474 | struct bin_tree_t *right;
|
|---|
| 475 | struct bin_tree_t *first;
|
|---|
| 476 | struct bin_tree_t *next;
|
|---|
| 477 |
|
|---|
| 478 | re_token_t token;
|
|---|
| 479 |
|
|---|
| 480 | /* 'node_idx' is the index in dfa->nodes, if 'type' == 0.
|
|---|
| 481 | Otherwise 'type' indicate the type of this node. */
|
|---|
| 482 | Idx node_idx;
|
|---|
| 483 | };
|
|---|
| 484 | typedef struct bin_tree_t bin_tree_t;
|
|---|
| 485 |
|
|---|
| 486 | #define BIN_TREE_STORAGE_SIZE \
|
|---|
| 487 | ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
|
|---|
| 488 |
|
|---|
| 489 | struct bin_tree_storage_t
|
|---|
| 490 | {
|
|---|
| 491 | struct bin_tree_storage_t *next;
|
|---|
| 492 | bin_tree_t data[BIN_TREE_STORAGE_SIZE];
|
|---|
| 493 | };
|
|---|
| 494 | typedef struct bin_tree_storage_t bin_tree_storage_t;
|
|---|
| 495 |
|
|---|
| 496 | #define CONTEXT_WORD 1
|
|---|
| 497 | #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
|
|---|
| 498 | #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
|
|---|
| 499 | #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
|
|---|
| 500 |
|
|---|
| 501 | #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
|
|---|
| 502 | #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
|
|---|
| 503 | #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
|
|---|
| 504 | #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
|
|---|
| 505 | #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
|
|---|
| 506 |
|
|---|
| 507 | #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
|
|---|
| 508 | #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
|
|---|
| 509 | #define IS_WIDE_WORD_CHAR(ch) (__iswalnum (ch) || (ch) == L'_')
|
|---|
| 510 | #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
|
|---|
| 511 |
|
|---|
| 512 | #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
|
|---|
| 513 | ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
|
|---|
| 514 | || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
|
|---|
| 515 | || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
|
|---|
| 516 | || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
|
|---|
| 517 |
|
|---|
| 518 | #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
|
|---|
| 519 | ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
|
|---|
| 520 | || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
|
|---|
| 521 | || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
|
|---|
| 522 | || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
|
|---|
| 523 |
|
|---|
| 524 | struct re_dfastate_t
|
|---|
| 525 | {
|
|---|
| 526 | re_hashval_t hash;
|
|---|
| 527 | re_node_set nodes;
|
|---|
| 528 | re_node_set non_eps_nodes;
|
|---|
| 529 | re_node_set inveclosure;
|
|---|
| 530 | re_node_set *entrance_nodes;
|
|---|
| 531 | struct re_dfastate_t **trtable, **word_trtable;
|
|---|
| 532 | unsigned int context : 4;
|
|---|
| 533 | unsigned int halt : 1;
|
|---|
| 534 | /* If this state can accept "multi byte".
|
|---|
| 535 | Note that we refer to multibyte characters, and multi character
|
|---|
| 536 | collating elements as "multi byte". */
|
|---|
| 537 | unsigned int accept_mb : 1;
|
|---|
| 538 | /* If this state has backreference node(s). */
|
|---|
| 539 | unsigned int has_backref : 1;
|
|---|
| 540 | unsigned int has_constraint : 1;
|
|---|
| 541 | };
|
|---|
| 542 | typedef struct re_dfastate_t re_dfastate_t;
|
|---|
| 543 |
|
|---|
| 544 | struct re_state_table_entry
|
|---|
| 545 | {
|
|---|
| 546 | Idx num;
|
|---|
| 547 | Idx alloc;
|
|---|
| 548 | re_dfastate_t **array;
|
|---|
| 549 | };
|
|---|
| 550 |
|
|---|
| 551 | /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
|
|---|
| 552 |
|
|---|
| 553 | typedef struct
|
|---|
| 554 | {
|
|---|
| 555 | Idx next_idx;
|
|---|
| 556 | Idx alloc;
|
|---|
| 557 | re_dfastate_t **array;
|
|---|
| 558 | } state_array_t;
|
|---|
| 559 |
|
|---|
| 560 | /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
|
|---|
| 561 |
|
|---|
| 562 | typedef struct
|
|---|
| 563 | {
|
|---|
| 564 | Idx node;
|
|---|
| 565 | Idx str_idx; /* The position NODE match at. */
|
|---|
| 566 | state_array_t path;
|
|---|
| 567 | } re_sub_match_last_t;
|
|---|
| 568 |
|
|---|
| 569 | /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
|
|---|
| 570 | And information about the node, whose type is OP_CLOSE_SUBEXP,
|
|---|
| 571 | corresponding to NODE is stored in LASTS. */
|
|---|
| 572 |
|
|---|
| 573 | typedef struct
|
|---|
| 574 | {
|
|---|
| 575 | Idx str_idx;
|
|---|
| 576 | Idx node;
|
|---|
| 577 | state_array_t *path;
|
|---|
| 578 | Idx alasts; /* Allocation size of LASTS. */
|
|---|
| 579 | Idx nlasts; /* The number of LASTS. */
|
|---|
| 580 | re_sub_match_last_t **lasts;
|
|---|
| 581 | } re_sub_match_top_t;
|
|---|
| 582 |
|
|---|
| 583 | struct re_backref_cache_entry
|
|---|
| 584 | {
|
|---|
| 585 | Idx node;
|
|---|
| 586 | Idx str_idx;
|
|---|
| 587 | Idx subexp_from;
|
|---|
| 588 | Idx subexp_to;
|
|---|
| 589 | bitset_word_t eps_reachable_subexps_map;
|
|---|
| 590 | char more;
|
|---|
| 591 | };
|
|---|
| 592 |
|
|---|
| 593 | typedef struct
|
|---|
| 594 | {
|
|---|
| 595 | /* The string object corresponding to the input string. */
|
|---|
| 596 | re_string_t input;
|
|---|
| 597 | const re_dfa_t *const dfa;
|
|---|
| 598 | /* EFLAGS of the argument of regexec. */
|
|---|
| 599 | int eflags;
|
|---|
| 600 | /* Where the matching ends. */
|
|---|
| 601 | Idx match_last;
|
|---|
| 602 | Idx last_node;
|
|---|
| 603 | /* The state log used by the matcher. */
|
|---|
| 604 | re_dfastate_t **state_log;
|
|---|
| 605 | Idx state_log_top;
|
|---|
| 606 | /* Back reference cache. */
|
|---|
| 607 | Idx nbkref_ents;
|
|---|
| 608 | Idx abkref_ents;
|
|---|
| 609 | struct re_backref_cache_entry *bkref_ents;
|
|---|
| 610 | int max_mb_elem_len;
|
|---|
| 611 | Idx nsub_tops;
|
|---|
| 612 | Idx asub_tops;
|
|---|
| 613 | re_sub_match_top_t **sub_tops;
|
|---|
| 614 | } re_match_context_t;
|
|---|
| 615 |
|
|---|
| 616 | typedef struct
|
|---|
| 617 | {
|
|---|
| 618 | re_dfastate_t **sifted_states;
|
|---|
| 619 | re_dfastate_t **limited_states;
|
|---|
| 620 | Idx last_node;
|
|---|
| 621 | Idx last_str_idx;
|
|---|
| 622 | re_node_set limits;
|
|---|
| 623 | } re_sift_context_t;
|
|---|
| 624 |
|
|---|
| 625 | struct re_fail_stack_ent_t
|
|---|
| 626 | {
|
|---|
| 627 | Idx idx;
|
|---|
| 628 | Idx node;
|
|---|
| 629 | regmatch_t *regs;
|
|---|
| 630 | re_node_set eps_via_nodes;
|
|---|
| 631 | };
|
|---|
| 632 |
|
|---|
| 633 | struct re_fail_stack_t
|
|---|
| 634 | {
|
|---|
| 635 | Idx num;
|
|---|
| 636 | Idx alloc;
|
|---|
| 637 | struct re_fail_stack_ent_t *stack;
|
|---|
| 638 | };
|
|---|
| 639 |
|
|---|
| 640 | struct re_dfa_t
|
|---|
| 641 | {
|
|---|
| 642 | re_token_t *nodes;
|
|---|
| 643 | size_t nodes_alloc;
|
|---|
| 644 | size_t nodes_len;
|
|---|
| 645 | Idx *nexts;
|
|---|
| 646 | Idx *org_indices;
|
|---|
| 647 | re_node_set *edests;
|
|---|
| 648 | re_node_set *eclosures;
|
|---|
| 649 | re_node_set *inveclosures;
|
|---|
| 650 | struct re_state_table_entry *state_table;
|
|---|
| 651 | re_dfastate_t *init_state;
|
|---|
| 652 | re_dfastate_t *init_state_word;
|
|---|
| 653 | re_dfastate_t *init_state_nl;
|
|---|
| 654 | re_dfastate_t *init_state_begbuf;
|
|---|
| 655 | bin_tree_t *str_tree;
|
|---|
| 656 | bin_tree_storage_t *str_tree_storage;
|
|---|
| 657 | re_bitset_ptr_t sb_char;
|
|---|
| 658 | int str_tree_storage_idx;
|
|---|
| 659 |
|
|---|
| 660 | /* number of subexpressions 're_nsub' is in regex_t. */
|
|---|
| 661 | re_hashval_t state_hash_mask;
|
|---|
| 662 | Idx init_node;
|
|---|
| 663 | Idx nbackref; /* The number of backreference in this dfa. */
|
|---|
| 664 |
|
|---|
| 665 | /* Bitmap expressing which backreference is used. */
|
|---|
| 666 | bitset_word_t used_bkref_map;
|
|---|
| 667 | bitset_word_t completed_bkref_map;
|
|---|
| 668 |
|
|---|
| 669 | unsigned int has_plural_match : 1;
|
|---|
| 670 | /* If this dfa has "multibyte node", which is a backreference or
|
|---|
| 671 | a node which can accept multibyte character or multi character
|
|---|
| 672 | collating element. */
|
|---|
| 673 | unsigned int has_mb_node : 1;
|
|---|
| 674 | unsigned int is_utf8 : 1;
|
|---|
| 675 | unsigned int map_notascii : 1;
|
|---|
| 676 | unsigned int word_ops_used : 1;
|
|---|
| 677 | int mb_cur_max;
|
|---|
| 678 | bitset_t word_char;
|
|---|
| 679 | reg_syntax_t syntax;
|
|---|
| 680 | Idx *subexp_map;
|
|---|
| 681 | #ifdef DEBUG
|
|---|
| 682 | char* re_str;
|
|---|
| 683 | #endif
|
|---|
| 684 | lock_define (lock)
|
|---|
| 685 | };
|
|---|
| 686 |
|
|---|
| 687 | #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
|
|---|
| 688 | #define re_node_set_remove(set,id) \
|
|---|
| 689 | (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
|
|---|
| 690 | #define re_node_set_empty(p) ((p)->nelem = 0)
|
|---|
| 691 | #define re_node_set_free(set) re_free ((set)->elems)
|
|---|
| 692 | |
|---|
| 693 |
|
|---|
| 694 |
|
|---|
| 695 | typedef enum
|
|---|
| 696 | {
|
|---|
| 697 | SB_CHAR,
|
|---|
| 698 | MB_CHAR,
|
|---|
| 699 | EQUIV_CLASS,
|
|---|
| 700 | COLL_SYM,
|
|---|
| 701 | CHAR_CLASS
|
|---|
| 702 | } bracket_elem_type;
|
|---|
| 703 |
|
|---|
| 704 | typedef struct
|
|---|
| 705 | {
|
|---|
| 706 | bracket_elem_type type;
|
|---|
| 707 | union
|
|---|
| 708 | {
|
|---|
| 709 | unsigned char ch;
|
|---|
| 710 | unsigned char *name;
|
|---|
| 711 | wchar_t wch;
|
|---|
| 712 | } opr;
|
|---|
| 713 | } bracket_elem_t;
|
|---|
| 714 |
|
|---|
| 715 |
|
|---|
| 716 | /* Functions for bitset_t operation. */
|
|---|
| 717 |
|
|---|
| 718 | static inline void
|
|---|
| 719 | bitset_set (bitset_t set, Idx i)
|
|---|
| 720 | {
|
|---|
| 721 | set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
|
|---|
| 722 | }
|
|---|
| 723 |
|
|---|
| 724 | static inline void
|
|---|
| 725 | bitset_clear (bitset_t set, Idx i)
|
|---|
| 726 | {
|
|---|
| 727 | set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
|
|---|
| 728 | }
|
|---|
| 729 |
|
|---|
| 730 | static inline bool
|
|---|
| 731 | bitset_contain (const bitset_t set, Idx i)
|
|---|
| 732 | {
|
|---|
| 733 | return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
|
|---|
| 734 | }
|
|---|
| 735 |
|
|---|
| 736 | static inline void
|
|---|
| 737 | bitset_empty (bitset_t set)
|
|---|
| 738 | {
|
|---|
| 739 | memset (set, '\0', sizeof (bitset_t));
|
|---|
| 740 | }
|
|---|
| 741 |
|
|---|
| 742 | static inline void
|
|---|
| 743 | bitset_set_all (bitset_t set)
|
|---|
| 744 | {
|
|---|
| 745 | memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
|
|---|
| 746 | if (SBC_MAX % BITSET_WORD_BITS != 0)
|
|---|
| 747 | set[BITSET_WORDS - 1] =
|
|---|
| 748 | ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
|
|---|
| 749 | }
|
|---|
| 750 |
|
|---|
| 751 | static inline void
|
|---|
| 752 | bitset_copy (bitset_t dest, const bitset_t src)
|
|---|
| 753 | {
|
|---|
| 754 | memcpy (dest, src, sizeof (bitset_t));
|
|---|
| 755 | }
|
|---|
| 756 |
|
|---|
| 757 | static inline void
|
|---|
| 758 | bitset_not (bitset_t set)
|
|---|
| 759 | {
|
|---|
| 760 | int bitset_i;
|
|---|
| 761 | for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
|
|---|
| 762 | set[bitset_i] = ~set[bitset_i];
|
|---|
| 763 | if (SBC_MAX % BITSET_WORD_BITS != 0)
|
|---|
| 764 | set[BITSET_WORDS - 1] =
|
|---|
| 765 | ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
|
|---|
| 766 | & ~set[BITSET_WORDS - 1]);
|
|---|
| 767 | }
|
|---|
| 768 |
|
|---|
| 769 | static inline void
|
|---|
| 770 | bitset_merge (bitset_t dest, const bitset_t src)
|
|---|
| 771 | {
|
|---|
| 772 | int bitset_i;
|
|---|
| 773 | for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
|
|---|
| 774 | dest[bitset_i] |= src[bitset_i];
|
|---|
| 775 | }
|
|---|
| 776 |
|
|---|
| 777 | static inline void
|
|---|
| 778 | bitset_mask (bitset_t dest, const bitset_t src)
|
|---|
| 779 | {
|
|---|
| 780 | int bitset_i;
|
|---|
| 781 | for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
|
|---|
| 782 | dest[bitset_i] &= src[bitset_i];
|
|---|
| 783 | }
|
|---|
| 784 |
|
|---|
| 785 | #ifdef RE_ENABLE_I18N
|
|---|
| 786 | /* Functions for re_string. */
|
|---|
| 787 | static int
|
|---|
| 788 | __attribute__ ((pure, unused))
|
|---|
| 789 | re_string_char_size_at (const re_string_t *pstr, Idx idx)
|
|---|
| 790 | {
|
|---|
| 791 | int byte_idx;
|
|---|
| 792 | if (pstr->mb_cur_max == 1)
|
|---|
| 793 | return 1;
|
|---|
| 794 | for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
|
|---|
| 795 | if (pstr->wcs[idx + byte_idx] != WEOF)
|
|---|
| 796 | break;
|
|---|
| 797 | return byte_idx;
|
|---|
| 798 | }
|
|---|
| 799 |
|
|---|
| 800 | static wint_t
|
|---|
| 801 | __attribute__ ((pure, unused))
|
|---|
| 802 | re_string_wchar_at (const re_string_t *pstr, Idx idx)
|
|---|
| 803 | {
|
|---|
| 804 | if (pstr->mb_cur_max == 1)
|
|---|
| 805 | return (wint_t) pstr->mbs[idx];
|
|---|
| 806 | return (wint_t) pstr->wcs[idx];
|
|---|
| 807 | }
|
|---|
| 808 |
|
|---|
| 809 | # ifdef _LIBC
|
|---|
| 810 | # include <locale/weight.h>
|
|---|
| 811 | # endif
|
|---|
| 812 |
|
|---|
| 813 | static int
|
|---|
| 814 | __attribute__ ((pure, unused))
|
|---|
| 815 | re_string_elem_size_at (const re_string_t *pstr, Idx idx)
|
|---|
| 816 | {
|
|---|
| 817 | # ifdef _LIBC
|
|---|
| 818 | const unsigned char *p, *extra;
|
|---|
| 819 | const int32_t *table, *indirect;
|
|---|
| 820 | uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
|
|---|
| 821 |
|
|---|
| 822 | if (nrules != 0)
|
|---|
| 823 | {
|
|---|
| 824 | table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
|
|---|
| 825 | extra = (const unsigned char *)
|
|---|
| 826 | _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
|
|---|
| 827 | indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
|
|---|
| 828 | _NL_COLLATE_INDIRECTMB);
|
|---|
| 829 | p = pstr->mbs + idx;
|
|---|
| 830 | findidx (table, indirect, extra, &p, pstr->len - idx);
|
|---|
| 831 | return p - pstr->mbs - idx;
|
|---|
| 832 | }
|
|---|
| 833 | else
|
|---|
| 834 | # endif /* _LIBC */
|
|---|
| 835 | return 1;
|
|---|
| 836 | }
|
|---|
| 837 | #endif /* RE_ENABLE_I18N */
|
|---|
| 838 |
|
|---|
| 839 | #ifdef _LIBC
|
|---|
| 840 | # if __GNUC__ >= 7
|
|---|
| 841 | # define FALLTHROUGH __attribute__ ((__fallthrough__))
|
|---|
| 842 | # else
|
|---|
| 843 | # define FALLTHROUGH ((void) 0)
|
|---|
| 844 | # endif
|
|---|
| 845 | #else
|
|---|
| 846 | # include "attribute.h"
|
|---|
| 847 | #endif
|
|---|
| 848 |
|
|---|
| 849 | #endif /* _REGEX_INTERNAL_H */
|
|---|