1 | /* Extended regular expression matching and search library. |
2 | Copyright (C) 2002-2017 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 | <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 | #if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET || defined _LIBC |
30 | # include <langinfo.h> |
31 | #endif |
32 | #if defined HAVE_LOCALE_H || defined _LIBC |
33 | # include <locale.h> |
34 | #endif |
35 | #if defined HAVE_WCHAR_H || defined _LIBC |
36 | # include <wchar.h> |
37 | #endif /* HAVE_WCHAR_H || _LIBC */ |
38 | #if defined HAVE_WCTYPE_H || defined _LIBC |
39 | # include <wctype.h> |
40 | #endif /* HAVE_WCTYPE_H || _LIBC */ |
41 | #if defined HAVE_STDBOOL_H || defined _LIBC |
42 | # include <stdbool.h> |
43 | #endif /* HAVE_STDBOOL_H || _LIBC */ |
44 | #if defined HAVE_STDINT_H || defined _LIBC |
45 | # include <stdint.h> |
46 | #endif /* HAVE_STDINT_H || _LIBC */ |
47 | #if defined _LIBC |
48 | # include <libc-lock.h> |
49 | #else |
50 | # define __libc_lock_define(CLASS,NAME) |
51 | # define __libc_lock_init(NAME) do { } while (0) |
52 | # define __libc_lock_lock(NAME) do { } while (0) |
53 | # define __libc_lock_unlock(NAME) do { } while (0) |
54 | #endif |
55 | |
56 | /* In case that the system doesn't have isblank(). */ |
57 | #if !defined _LIBC && !defined HAVE_ISBLANK && !defined isblank |
58 | # define isblank(ch) ((ch) == ' ' || (ch) == '\t') |
59 | #endif |
60 | |
61 | #ifdef _LIBC |
62 | # ifndef _RE_DEFINE_LOCALE_FUNCTIONS |
63 | # define _RE_DEFINE_LOCALE_FUNCTIONS 1 |
64 | # include <locale/localeinfo.h> |
65 | # include <locale/coll-lookup.h> |
66 | # endif |
67 | #endif |
68 | |
69 | /* This is for other GNU distributions with internationalized messages. */ |
70 | #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC |
71 | # include <libintl.h> |
72 | # ifdef _LIBC |
73 | # undef gettext |
74 | # define gettext(msgid) \ |
75 | __dcgettext (_libc_intl_domainname, msgid, LC_MESSAGES) |
76 | # endif |
77 | #else |
78 | # define gettext(msgid) (msgid) |
79 | #endif |
80 | |
81 | #ifndef gettext_noop |
82 | /* This define is so xgettext can find the internationalizable |
83 | strings. */ |
84 | # define gettext_noop(String) String |
85 | #endif |
86 | |
87 | /* For loser systems without the definition. */ |
88 | #ifndef SIZE_MAX |
89 | # define SIZE_MAX ((size_t) -1) |
90 | #endif |
91 | |
92 | #if (defined MB_CUR_MAX && HAVE_WCTYPE_H && HAVE_ISWCTYPE) || _LIBC |
93 | # define RE_ENABLE_I18N |
94 | #endif |
95 | |
96 | #if __GNUC__ >= 3 |
97 | # define BE(expr, val) __builtin_expect (expr, val) |
98 | #else |
99 | # define BE(expr, val) (expr) |
100 | #endif |
101 | |
102 | /* Number of single byte character. */ |
103 | #define SBC_MAX 256 |
104 | |
105 | #define COLL_ELEM_LEN_MAX 8 |
106 | |
107 | /* The character which represents newline. */ |
108 | #define NEWLINE_CHAR '\n' |
109 | #define WIDE_NEWLINE_CHAR L'\n' |
110 | |
111 | /* Rename to standard API for using out of glibc. */ |
112 | #ifndef _LIBC |
113 | # define __wctype wctype |
114 | # define __iswctype iswctype |
115 | # define __btowc btowc |
116 | # define __mbrtowc mbrtowc |
117 | # define __mempcpy mempcpy |
118 | # define __wcrtomb wcrtomb |
119 | # define __regfree regfree |
120 | # define attribute_hidden |
121 | #endif /* not _LIBC */ |
122 | |
123 | #if __GNUC__ < 3 + (__GNUC_MINOR__ < 1) |
124 | # define __attribute__(arg) |
125 | #endif |
126 | |
127 | extern const char __re_error_msgid[] attribute_hidden; |
128 | extern const size_t __re_error_msgid_idx[] attribute_hidden; |
129 | |
130 | /* An integer used to represent a set of bits. It must be unsigned, |
131 | and must be at least as wide as unsigned int. */ |
132 | typedef unsigned long int bitset_word_t; |
133 | /* All bits set in a bitset_word_t. */ |
134 | #define BITSET_WORD_MAX ULONG_MAX |
135 | /* Number of bits in a bitset_word_t. */ |
136 | #define BITSET_WORD_BITS (sizeof (bitset_word_t) * CHAR_BIT) |
137 | /* Number of bitset_word_t in a bit_set. */ |
138 | #define BITSET_WORDS (SBC_MAX / BITSET_WORD_BITS) |
139 | typedef bitset_word_t bitset_t[BITSET_WORDS]; |
140 | typedef bitset_word_t *re_bitset_ptr_t; |
141 | typedef const bitset_word_t *re_const_bitset_ptr_t; |
142 | |
143 | #define bitset_set(set,i) \ |
144 | (set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS) |
145 | #define bitset_clear(set,i) \ |
146 | (set[i / BITSET_WORD_BITS] &= ~((bitset_word_t) 1 << i % BITSET_WORD_BITS)) |
147 | #define bitset_contain(set,i) \ |
148 | (set[i / BITSET_WORD_BITS] & ((bitset_word_t) 1 << i % BITSET_WORD_BITS)) |
149 | #define bitset_empty(set) memset (set, '\0', sizeof (bitset_t)) |
150 | #define bitset_set_all(set) memset (set, '\xff', sizeof (bitset_t)) |
151 | #define bitset_copy(dest,src) memcpy (dest, src, sizeof (bitset_t)) |
152 | |
153 | #define PREV_WORD_CONSTRAINT 0x0001 |
154 | #define PREV_NOTWORD_CONSTRAINT 0x0002 |
155 | #define NEXT_WORD_CONSTRAINT 0x0004 |
156 | #define NEXT_NOTWORD_CONSTRAINT 0x0008 |
157 | #define PREV_NEWLINE_CONSTRAINT 0x0010 |
158 | #define NEXT_NEWLINE_CONSTRAINT 0x0020 |
159 | #define PREV_BEGBUF_CONSTRAINT 0x0040 |
160 | #define NEXT_ENDBUF_CONSTRAINT 0x0080 |
161 | #define WORD_DELIM_CONSTRAINT 0x0100 |
162 | #define NOT_WORD_DELIM_CONSTRAINT 0x0200 |
163 | |
164 | typedef enum |
165 | { |
166 | INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, |
167 | WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT, |
168 | WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT, |
169 | INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT, |
170 | LINE_FIRST = PREV_NEWLINE_CONSTRAINT, |
171 | LINE_LAST = NEXT_NEWLINE_CONSTRAINT, |
172 | BUF_FIRST = PREV_BEGBUF_CONSTRAINT, |
173 | BUF_LAST = NEXT_ENDBUF_CONSTRAINT, |
174 | WORD_DELIM = WORD_DELIM_CONSTRAINT, |
175 | NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT |
176 | } re_context_type; |
177 | |
178 | typedef struct |
179 | { |
180 | int alloc; |
181 | int nelem; |
182 | int *elems; |
183 | } re_node_set; |
184 | |
185 | typedef enum |
186 | { |
187 | NON_TYPE = 0, |
188 | |
189 | /* Node type, These are used by token, node, tree. */ |
190 | CHARACTER = 1, |
191 | END_OF_RE = 2, |
192 | SIMPLE_BRACKET = 3, |
193 | OP_BACK_REF = 4, |
194 | OP_PERIOD = 5, |
195 | #ifdef RE_ENABLE_I18N |
196 | COMPLEX_BRACKET = 6, |
197 | OP_UTF8_PERIOD = 7, |
198 | #endif /* RE_ENABLE_I18N */ |
199 | |
200 | /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used |
201 | when the debugger shows values of this enum type. */ |
202 | #define EPSILON_BIT 8 |
203 | OP_OPEN_SUBEXP = EPSILON_BIT | 0, |
204 | OP_CLOSE_SUBEXP = EPSILON_BIT | 1, |
205 | OP_ALT = EPSILON_BIT | 2, |
206 | OP_DUP_ASTERISK = EPSILON_BIT | 3, |
207 | ANCHOR = EPSILON_BIT | 4, |
208 | |
209 | /* Tree type, these are used only by tree. */ |
210 | CONCAT = 16, |
211 | SUBEXP = 17, |
212 | |
213 | /* Token type, these are used only by token. */ |
214 | OP_DUP_PLUS = 18, |
215 | OP_DUP_QUESTION, |
216 | OP_OPEN_BRACKET, |
217 | OP_CLOSE_BRACKET, |
218 | OP_CHARSET_RANGE, |
219 | OP_OPEN_DUP_NUM, |
220 | OP_CLOSE_DUP_NUM, |
221 | OP_NON_MATCH_LIST, |
222 | OP_OPEN_COLL_ELEM, |
223 | OP_CLOSE_COLL_ELEM, |
224 | OP_OPEN_EQUIV_CLASS, |
225 | OP_CLOSE_EQUIV_CLASS, |
226 | OP_OPEN_CHAR_CLASS, |
227 | OP_CLOSE_CHAR_CLASS, |
228 | OP_WORD, |
229 | OP_NOTWORD, |
230 | OP_SPACE, |
231 | OP_NOTSPACE, |
232 | BACK_SLASH |
233 | |
234 | } re_token_type_t; |
235 | |
236 | #ifdef RE_ENABLE_I18N |
237 | typedef struct |
238 | { |
239 | /* Multibyte characters. */ |
240 | wchar_t *mbchars; |
241 | |
242 | /* Collating symbols. */ |
243 | # ifdef _LIBC |
244 | int32_t *coll_syms; |
245 | # endif |
246 | |
247 | /* Equivalence classes. */ |
248 | # ifdef _LIBC |
249 | int32_t *equiv_classes; |
250 | # endif |
251 | |
252 | /* Range expressions. */ |
253 | # ifdef _LIBC |
254 | uint32_t *range_starts; |
255 | uint32_t *range_ends; |
256 | # else /* not _LIBC */ |
257 | wchar_t *range_starts; |
258 | wchar_t *range_ends; |
259 | # endif /* not _LIBC */ |
260 | |
261 | /* Character classes. */ |
262 | wctype_t *char_classes; |
263 | |
264 | /* If this character set is the non-matching list. */ |
265 | unsigned int non_match : 1; |
266 | |
267 | /* # of multibyte characters. */ |
268 | int nmbchars; |
269 | |
270 | /* # of collating symbols. */ |
271 | int ncoll_syms; |
272 | |
273 | /* # of equivalence classes. */ |
274 | int nequiv_classes; |
275 | |
276 | /* # of range expressions. */ |
277 | int nranges; |
278 | |
279 | /* # of character classes. */ |
280 | int nchar_classes; |
281 | } re_charset_t; |
282 | #endif /* RE_ENABLE_I18N */ |
283 | |
284 | typedef struct |
285 | { |
286 | union |
287 | { |
288 | unsigned char c; /* for CHARACTER */ |
289 | re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */ |
290 | #ifdef RE_ENABLE_I18N |
291 | re_charset_t *mbcset; /* for COMPLEX_BRACKET */ |
292 | #endif /* RE_ENABLE_I18N */ |
293 | int idx; /* for BACK_REF */ |
294 | re_context_type ctx_type; /* for ANCHOR */ |
295 | } opr; |
296 | #if __GNUC__ >= 2 |
297 | re_token_type_t type : 8; |
298 | #else |
299 | re_token_type_t type; |
300 | #endif |
301 | unsigned int constraint : 10; /* context constraint */ |
302 | unsigned int duplicated : 1; |
303 | unsigned int opt_subexp : 1; |
304 | #ifdef RE_ENABLE_I18N |
305 | unsigned int accept_mb : 1; |
306 | /* These 2 bits can be moved into the union if needed (e.g. if running out |
307 | of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */ |
308 | unsigned int mb_partial : 1; |
309 | #endif |
310 | unsigned int word_char : 1; |
311 | } re_token_t; |
312 | |
313 | #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT) |
314 | |
315 | struct re_string_t |
316 | { |
317 | /* Indicate the raw buffer which is the original string passed as an |
318 | argument of regexec(), re_search(), etc.. */ |
319 | const unsigned char *raw_mbs; |
320 | /* Store the multibyte string. In case of "case insensitive mode" like |
321 | REG_ICASE, upper cases of the string are stored, otherwise MBS points |
322 | the same address that RAW_MBS points. */ |
323 | unsigned char *mbs; |
324 | #ifdef RE_ENABLE_I18N |
325 | /* Store the wide character string which is corresponding to MBS. */ |
326 | wint_t *wcs; |
327 | int *offsets; |
328 | mbstate_t cur_state; |
329 | #endif |
330 | /* Index in RAW_MBS. Each character mbs[i] corresponds to |
331 | raw_mbs[raw_mbs_idx + i]. */ |
332 | int raw_mbs_idx; |
333 | /* The length of the valid characters in the buffers. */ |
334 | int valid_len; |
335 | /* The corresponding number of bytes in raw_mbs array. */ |
336 | int valid_raw_len; |
337 | /* The length of the buffers MBS and WCS. */ |
338 | int bufs_len; |
339 | /* The index in MBS, which is updated by re_string_fetch_byte. */ |
340 | int cur_idx; |
341 | /* length of RAW_MBS array. */ |
342 | int raw_len; |
343 | /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */ |
344 | int len; |
345 | /* End of the buffer may be shorter than its length in the cases such |
346 | as re_match_2, re_search_2. Then, we use STOP for end of the buffer |
347 | instead of LEN. */ |
348 | int raw_stop; |
349 | /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */ |
350 | int stop; |
351 | |
352 | /* The context of mbs[0]. We store the context independently, since |
353 | the context of mbs[0] may be different from raw_mbs[0], which is |
354 | the beginning of the input string. */ |
355 | unsigned int tip_context; |
356 | /* The translation passed as a part of an argument of re_compile_pattern. */ |
357 | RE_TRANSLATE_TYPE trans; |
358 | /* Copy of re_dfa_t's word_char. */ |
359 | re_const_bitset_ptr_t word_char; |
360 | /* 1 if REG_ICASE. */ |
361 | unsigned char icase; |
362 | unsigned char is_utf8; |
363 | unsigned char map_notascii; |
364 | unsigned char mbs_allocated; |
365 | unsigned char offsets_needed; |
366 | unsigned char newline_anchor; |
367 | unsigned char word_ops_used; |
368 | int mb_cur_max; |
369 | }; |
370 | typedef struct re_string_t re_string_t; |
371 | |
372 | |
373 | struct re_dfa_t; |
374 | typedef struct re_dfa_t re_dfa_t; |
375 | |
376 | #ifndef _LIBC |
377 | # ifdef __i386__ |
378 | # define internal_function __attribute__ ((regparm (3), stdcall)) |
379 | # else |
380 | # define internal_function |
381 | # endif |
382 | #endif |
383 | |
384 | #if IS_IN (libc) |
385 | static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr, |
386 | int new_buf_len) |
387 | internal_function; |
388 | # ifdef RE_ENABLE_I18N |
389 | static void build_wcs_buffer (re_string_t *pstr) internal_function; |
390 | static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr) |
391 | internal_function; |
392 | # endif /* RE_ENABLE_I18N */ |
393 | static void build_upper_buffer (re_string_t *pstr) internal_function; |
394 | static void re_string_translate_buffer (re_string_t *pstr) internal_function; |
395 | static unsigned int re_string_context_at (const re_string_t *input, int idx, |
396 | int eflags) |
397 | internal_function __attribute__ ((pure)); |
398 | #endif |
399 | #define re_string_peek_byte(pstr, offset) \ |
400 | ((pstr)->mbs[(pstr)->cur_idx + offset]) |
401 | #define re_string_fetch_byte(pstr) \ |
402 | ((pstr)->mbs[(pstr)->cur_idx++]) |
403 | #define re_string_first_byte(pstr, idx) \ |
404 | ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF) |
405 | #define re_string_is_single_byte_char(pstr, idx) \ |
406 | ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \ |
407 | || (pstr)->wcs[(idx) + 1] != WEOF)) |
408 | #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx) |
409 | #define re_string_cur_idx(pstr) ((pstr)->cur_idx) |
410 | #define re_string_get_buffer(pstr) ((pstr)->mbs) |
411 | #define re_string_length(pstr) ((pstr)->len) |
412 | #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx]) |
413 | #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx)) |
414 | #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx)) |
415 | |
416 | #include <alloca.h> |
417 | |
418 | #ifndef _LIBC |
419 | # if HAVE_ALLOCA |
420 | /* The OS usually guarantees only one guard page at the bottom of the stack, |
421 | and a page size can be as small as 4096 bytes. So we cannot safely |
422 | allocate anything larger than 4096 bytes. Also care for the possibility |
423 | of a few compiler-allocated temporary stack slots. */ |
424 | # define __libc_use_alloca(n) ((n) < 4032) |
425 | # else |
426 | /* alloca is implemented with malloc, so just use malloc. */ |
427 | # define __libc_use_alloca(n) 0 |
428 | # endif |
429 | #endif |
430 | |
431 | #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t))) |
432 | #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t))) |
433 | #define re_free(p) free (p) |
434 | |
435 | struct bin_tree_t |
436 | { |
437 | struct bin_tree_t *parent; |
438 | struct bin_tree_t *left; |
439 | struct bin_tree_t *right; |
440 | struct bin_tree_t *first; |
441 | struct bin_tree_t *next; |
442 | |
443 | re_token_t token; |
444 | |
445 | /* `node_idx' is the index in dfa->nodes, if `type' == 0. |
446 | Otherwise `type' indicate the type of this node. */ |
447 | int node_idx; |
448 | }; |
449 | typedef struct bin_tree_t bin_tree_t; |
450 | |
451 | #define BIN_TREE_STORAGE_SIZE \ |
452 | ((1024 - sizeof (void *)) / sizeof (bin_tree_t)) |
453 | |
454 | struct bin_tree_storage_t |
455 | { |
456 | struct bin_tree_storage_t *next; |
457 | bin_tree_t data[BIN_TREE_STORAGE_SIZE]; |
458 | }; |
459 | typedef struct bin_tree_storage_t bin_tree_storage_t; |
460 | |
461 | #define CONTEXT_WORD 1 |
462 | #define CONTEXT_NEWLINE (CONTEXT_WORD << 1) |
463 | #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1) |
464 | #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1) |
465 | |
466 | #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD) |
467 | #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE) |
468 | #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF) |
469 | #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF) |
470 | #define IS_ORDINARY_CONTEXT(c) ((c) == 0) |
471 | |
472 | #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_') |
473 | #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR) |
474 | #define IS_WIDE_WORD_CHAR(ch) (__iswalnum (ch) || (ch) == L'_') |
475 | #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR) |
476 | |
477 | #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \ |
478 | ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ |
479 | || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ |
480 | || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\ |
481 | || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context))) |
482 | |
483 | #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \ |
484 | ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \ |
485 | || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \ |
486 | || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \ |
487 | || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context))) |
488 | |
489 | struct re_dfastate_t |
490 | { |
491 | unsigned int hash; |
492 | re_node_set nodes; |
493 | re_node_set non_eps_nodes; |
494 | re_node_set inveclosure; |
495 | re_node_set *entrance_nodes; |
496 | struct re_dfastate_t **trtable, **word_trtable; |
497 | unsigned int context : 4; |
498 | unsigned int halt : 1; |
499 | /* If this state can accept `multi byte'. |
500 | Note that we refer to multibyte characters, and multi character |
501 | collating elements as `multi byte'. */ |
502 | unsigned int accept_mb : 1; |
503 | /* If this state has backreference node(s). */ |
504 | unsigned int has_backref : 1; |
505 | unsigned int has_constraint : 1; |
506 | }; |
507 | typedef struct re_dfastate_t re_dfastate_t; |
508 | |
509 | struct re_state_table_entry |
510 | { |
511 | int num; |
512 | int alloc; |
513 | re_dfastate_t **array; |
514 | }; |
515 | |
516 | /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */ |
517 | |
518 | typedef struct |
519 | { |
520 | int next_idx; |
521 | int alloc; |
522 | re_dfastate_t **array; |
523 | } state_array_t; |
524 | |
525 | /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */ |
526 | |
527 | typedef struct |
528 | { |
529 | int node; |
530 | int str_idx; /* The position NODE match at. */ |
531 | state_array_t path; |
532 | } re_sub_match_last_t; |
533 | |
534 | /* Store information about the node NODE whose type is OP_OPEN_SUBEXP. |
535 | And information about the node, whose type is OP_CLOSE_SUBEXP, |
536 | corresponding to NODE is stored in LASTS. */ |
537 | |
538 | typedef struct |
539 | { |
540 | int str_idx; |
541 | int node; |
542 | state_array_t *path; |
543 | int alasts; /* Allocation size of LASTS. */ |
544 | int nlasts; /* The number of LASTS. */ |
545 | re_sub_match_last_t **lasts; |
546 | } re_sub_match_top_t; |
547 | |
548 | struct re_backref_cache_entry |
549 | { |
550 | int node; |
551 | int str_idx; |
552 | int subexp_from; |
553 | int subexp_to; |
554 | char more; |
555 | char unused; |
556 | unsigned short int eps_reachable_subexps_map; |
557 | }; |
558 | |
559 | typedef struct |
560 | { |
561 | /* The string object corresponding to the input string. */ |
562 | re_string_t input; |
563 | #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L) |
564 | const re_dfa_t *const dfa; |
565 | #else |
566 | const re_dfa_t *dfa; |
567 | #endif |
568 | /* EFLAGS of the argument of regexec. */ |
569 | int eflags; |
570 | /* Where the matching ends. */ |
571 | int match_last; |
572 | int last_node; |
573 | /* The state log used by the matcher. */ |
574 | re_dfastate_t **state_log; |
575 | int state_log_top; |
576 | /* Back reference cache. */ |
577 | int nbkref_ents; |
578 | int abkref_ents; |
579 | struct re_backref_cache_entry *bkref_ents; |
580 | int max_mb_elem_len; |
581 | int nsub_tops; |
582 | int asub_tops; |
583 | re_sub_match_top_t **sub_tops; |
584 | } re_match_context_t; |
585 | |
586 | typedef struct |
587 | { |
588 | re_dfastate_t **sifted_states; |
589 | re_dfastate_t **limited_states; |
590 | int last_node; |
591 | int last_str_idx; |
592 | re_node_set limits; |
593 | } re_sift_context_t; |
594 | |
595 | struct re_fail_stack_ent_t |
596 | { |
597 | int idx; |
598 | int node; |
599 | regmatch_t *regs; |
600 | re_node_set eps_via_nodes; |
601 | }; |
602 | |
603 | struct re_fail_stack_t |
604 | { |
605 | int num; |
606 | int alloc; |
607 | struct re_fail_stack_ent_t *stack; |
608 | }; |
609 | |
610 | struct re_dfa_t |
611 | { |
612 | re_token_t *nodes; |
613 | size_t nodes_alloc; |
614 | size_t nodes_len; |
615 | int *nexts; |
616 | int *org_indices; |
617 | re_node_set *edests; |
618 | re_node_set *eclosures; |
619 | re_node_set *inveclosures; |
620 | struct re_state_table_entry *state_table; |
621 | re_dfastate_t *init_state; |
622 | re_dfastate_t *init_state_word; |
623 | re_dfastate_t *init_state_nl; |
624 | re_dfastate_t *init_state_begbuf; |
625 | bin_tree_t *str_tree; |
626 | bin_tree_storage_t *str_tree_storage; |
627 | re_bitset_ptr_t sb_char; |
628 | int str_tree_storage_idx; |
629 | |
630 | /* number of subexpressions `re_nsub' is in regex_t. */ |
631 | unsigned int state_hash_mask; |
632 | int init_node; |
633 | int nbackref; /* The number of backreference in this dfa. */ |
634 | |
635 | /* Bitmap expressing which backreference is used. */ |
636 | bitset_word_t used_bkref_map; |
637 | bitset_word_t completed_bkref_map; |
638 | |
639 | unsigned int has_plural_match : 1; |
640 | /* If this dfa has "multibyte node", which is a backreference or |
641 | a node which can accept multibyte character or multi character |
642 | collating element. */ |
643 | unsigned int has_mb_node : 1; |
644 | unsigned int is_utf8 : 1; |
645 | unsigned int map_notascii : 1; |
646 | unsigned int word_ops_used : 1; |
647 | int mb_cur_max; |
648 | bitset_t word_char; |
649 | reg_syntax_t syntax; |
650 | int *subexp_map; |
651 | #ifdef DEBUG |
652 | char* re_str; |
653 | #endif |
654 | __libc_lock_define (, lock) |
655 | }; |
656 | |
657 | #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set)) |
658 | #define re_node_set_remove(set,id) \ |
659 | (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1)) |
660 | #define re_node_set_empty(p) ((p)->nelem = 0) |
661 | #define re_node_set_free(set) re_free ((set)->elems) |
662 | |
663 | |
664 | typedef enum |
665 | { |
666 | SB_CHAR, |
667 | MB_CHAR, |
668 | EQUIV_CLASS, |
669 | COLL_SYM, |
670 | CHAR_CLASS |
671 | } bracket_elem_type; |
672 | |
673 | typedef struct |
674 | { |
675 | bracket_elem_type type; |
676 | union |
677 | { |
678 | unsigned char ch; |
679 | unsigned char *name; |
680 | wchar_t wch; |
681 | } opr; |
682 | } bracket_elem_t; |
683 | |
684 | |
685 | /* Inline functions for bitset operation. */ |
686 | static void __attribute__ ((unused)) |
687 | bitset_not (bitset_t set) |
688 | { |
689 | int bitset_i; |
690 | for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i) |
691 | set[bitset_i] = ~set[bitset_i]; |
692 | } |
693 | |
694 | static void __attribute__ ((unused)) |
695 | bitset_merge (bitset_t dest, const bitset_t src) |
696 | { |
697 | int bitset_i; |
698 | for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i) |
699 | dest[bitset_i] |= src[bitset_i]; |
700 | } |
701 | |
702 | static void __attribute__ ((unused)) |
703 | bitset_mask (bitset_t dest, const bitset_t src) |
704 | { |
705 | int bitset_i; |
706 | for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i) |
707 | dest[bitset_i] &= src[bitset_i]; |
708 | } |
709 | |
710 | #ifdef RE_ENABLE_I18N |
711 | /* Inline functions for re_string. */ |
712 | static int |
713 | internal_function __attribute__ ((pure, unused)) |
714 | re_string_char_size_at (const re_string_t *pstr, int idx) |
715 | { |
716 | int byte_idx; |
717 | if (pstr->mb_cur_max == 1) |
718 | return 1; |
719 | for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx) |
720 | if (pstr->wcs[idx + byte_idx] != WEOF) |
721 | break; |
722 | return byte_idx; |
723 | } |
724 | |
725 | static wint_t |
726 | internal_function __attribute__ ((pure, unused)) |
727 | re_string_wchar_at (const re_string_t *pstr, int idx) |
728 | { |
729 | if (pstr->mb_cur_max == 1) |
730 | return (wint_t) pstr->mbs[idx]; |
731 | return (wint_t) pstr->wcs[idx]; |
732 | } |
733 | |
734 | # if IS_IN (libc) |
735 | # ifdef _LIBC |
736 | # include <locale/weight.h> |
737 | # endif |
738 | |
739 | static int |
740 | internal_function __attribute__ ((pure, unused)) |
741 | re_string_elem_size_at (const re_string_t *pstr, int idx) |
742 | { |
743 | # ifdef _LIBC |
744 | const unsigned char *p, *; |
745 | const int32_t *table, *indirect; |
746 | uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES); |
747 | |
748 | if (nrules != 0) |
749 | { |
750 | table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB); |
751 | extra = (const unsigned char *) |
752 | _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB); |
753 | indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE, |
754 | _NL_COLLATE_INDIRECTMB); |
755 | p = pstr->mbs + idx; |
756 | findidx (table, indirect, extra, &p, pstr->len - idx); |
757 | return p - pstr->mbs - idx; |
758 | } |
759 | else |
760 | # endif /* _LIBC */ |
761 | return 1; |
762 | } |
763 | # endif |
764 | #endif /* RE_ENABLE_I18N */ |
765 | |
766 | #endif /* _REGEX_INTERNAL_H */ |
767 | |