Line data Source code
1 : /* -*- buffer-read-only: t -*- vi: set ro: */
2 : /* DO NOT EDIT! GENERATED AUTOMATICALLY! */
3 : #line 1
4 : /* Extended regular expression matching and search library.
5 : Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc.
6 : This file is part of the GNU C Library.
7 : Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
8 :
9 : This program is free software; you can redistribute it and/or modify
10 : it under the terms of the GNU General Public License as published by
11 : the Free Software Foundation; either version 3, or (at your option)
12 : any later version.
13 :
14 : This program is distributed in the hope that it will be useful,
15 : but WITHOUT ANY WARRANTY; without even the implied warranty of
16 : MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 : GNU General Public License for more details.
18 :
19 : You should have received a copy of the GNU General Public License along
20 : with this program; if not, write to the Free Software Foundation,
21 : Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
22 :
23 : #ifndef _REGEX_INTERNAL_H
24 : #define _REGEX_INTERNAL_H 1
25 :
26 : #include <assert.h>
27 : #include <ctype.h>
28 : #include <stdbool.h>
29 : #include <stdio.h>
30 : #include <stdlib.h>
31 : #include <string.h>
32 :
33 : #ifdef _LIBC
34 : # include <langinfo.h>
35 : #else
36 : # include "localcharset.h"
37 : #endif
38 : #if defined HAVE_LOCALE_H || defined _LIBC
39 : # include <locale.h>
40 : #endif
41 :
42 : #include <wchar.h>
43 : #include <wctype.h>
44 : #include <stdint.h>
45 : #if defined _LIBC
46 : # include <bits/libc-lock.h>
47 : #else
48 : # define __libc_lock_init(NAME) do { } while (0)
49 : # define __libc_lock_lock(NAME) do { } while (0)
50 : # define __libc_lock_unlock(NAME) do { } while (0)
51 : #endif
52 :
53 : /* In case that the system doesn't have isblank(). */
54 : #if !defined _LIBC && ! (defined isblank || (HAVE_ISBLANK && HAVE_DECL_ISBLANK))
55 : # define isblank(ch) ((ch) == ' ' || (ch) == '\t')
56 : #endif
57 :
58 : #ifdef _LIBC
59 : # ifndef _RE_DEFINE_LOCALE_FUNCTIONS
60 : # define _RE_DEFINE_LOCALE_FUNCTIONS 1
61 : # include <locale/localeinfo.h>
62 : # include <locale/elem-hash.h>
63 : # include <locale/coll-lookup.h>
64 : # endif
65 : #endif
66 :
67 : /* This is for other GNU distributions with internationalized messages. */
68 : #if (HAVE_LIBINTL_H && ENABLE_NLS) || defined _LIBC
69 : # include <libintl.h>
70 : # ifdef _LIBC
71 : # undef gettext
72 : # define gettext(msgid) \
73 : INTUSE(__dcgettext) (_libc_intl_domainname, msgid, LC_MESSAGES)
74 : # endif
75 : #else
76 : # define gettext(msgid) (msgid)
77 : #endif
78 :
79 : #ifndef gettext_noop
80 : /* This define is so xgettext can find the internationalizable
81 : strings. */
82 : # define gettext_noop(String) String
83 : #endif
84 :
85 : /* For loser systems without the definition. */
86 : #ifndef SIZE_MAX
87 : # define SIZE_MAX ((size_t) -1)
88 : #endif
89 :
90 : #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_ISWCTYPE && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL) || _LIBC
91 : # define RE_ENABLE_I18N
92 : #endif
93 :
94 : #if __GNUC__ >= 3
95 : # define BE(expr, val) __builtin_expect (expr, val)
96 : #else
97 : # define BE(expr, val) (expr)
98 : # ifdef _LIBC
99 : # define inline
100 : # endif
101 : #endif
102 :
103 : /* Number of ASCII characters. */
104 : #define ASCII_CHARS 0x80
105 :
106 : /* Number of single byte characters. */
107 : #define SBC_MAX (UCHAR_MAX + 1)
108 :
109 : #define COLL_ELEM_LEN_MAX 8
110 :
111 : /* The character which represents newline. */
112 : #define NEWLINE_CHAR '\n'
113 : #define WIDE_NEWLINE_CHAR L'\n'
114 :
115 : /* Rename to standard API for using out of glibc. */
116 : #ifndef _LIBC
117 : # define __wctype wctype
118 : # define __iswctype iswctype
119 : # define __btowc btowc
120 : # define __wcrtomb wcrtomb
121 : # define __regfree regfree
122 : # define attribute_hidden
123 : #endif /* not _LIBC */
124 :
125 : #if __GNUC__ >= 4 || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1)
126 : # define __attribute(arg) __attribute__ (arg)
127 : #else
128 : # define __attribute(arg)
129 : #endif
130 :
131 : typedef __re_idx_t Idx;
132 :
133 : /* Special return value for failure to match. */
134 : #define REG_MISSING ((Idx) -1)
135 :
136 : /* Special return value for internal error. */
137 : #define REG_ERROR ((Idx) -2)
138 :
139 : /* Test whether N is a valid index, and is not one of the above. */
140 : #ifdef _REGEX_LARGE_OFFSETS
141 : # define REG_VALID_INDEX(n) ((Idx) (n) < REG_ERROR)
142 : #else
143 : # define REG_VALID_INDEX(n) (0 <= (n))
144 : #endif
145 :
146 : /* Test whether N is a valid nonzero index. */
147 : #ifdef _REGEX_LARGE_OFFSETS
148 : # define REG_VALID_NONZERO_INDEX(n) ((Idx) ((n) - 1) < (Idx) (REG_ERROR - 1))
149 : #else
150 : # define REG_VALID_NONZERO_INDEX(n) (0 < (n))
151 : #endif
152 :
153 : /* A hash value, suitable for computing hash tables. */
154 : typedef __re_size_t re_hashval_t;
155 :
156 : /* An integer used to represent a set of bits. It must be unsigned,
157 : and must be at least as wide as unsigned int. */
158 : typedef unsigned long int bitset_word_t;
159 : /* All bits set in a bitset_word_t. */
160 : #define BITSET_WORD_MAX ULONG_MAX
161 :
162 : /* Number of bits in a bitset_word_t. For portability to hosts with
163 : padding bits, do not use '(sizeof (bitset_word_t) * CHAR_BIT)';
164 : instead, deduce it directly from BITSET_WORD_MAX. Avoid
165 : greater-than-32-bit integers and unconditional shifts by more than
166 : 31 bits, as they're not portable. */
167 : #if BITSET_WORD_MAX == 0xffffffff
168 : # define BITSET_WORD_BITS 32
169 : #elif BITSET_WORD_MAX >> 31 >> 5 == 1
170 : # define BITSET_WORD_BITS 36
171 : #elif BITSET_WORD_MAX >> 31 >> 16 == 1
172 : # define BITSET_WORD_BITS 48
173 : #elif BITSET_WORD_MAX >> 31 >> 28 == 1
174 : # define BITSET_WORD_BITS 60
175 : #elif BITSET_WORD_MAX >> 31 >> 31 >> 1 == 1
176 : # define BITSET_WORD_BITS 64
177 : #elif BITSET_WORD_MAX >> 31 >> 31 >> 9 == 1
178 : # define BITSET_WORD_BITS 72
179 : #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 3 == 1
180 : # define BITSET_WORD_BITS 128
181 : #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 == 1
182 : # define BITSET_WORD_BITS 256
183 : #elif BITSET_WORD_MAX >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 31 >> 7 > 1
184 : # define BITSET_WORD_BITS 257 /* any value > SBC_MAX will do here */
185 : # if BITSET_WORD_BITS <= SBC_MAX
186 : # error "Invalid SBC_MAX"
187 : # endif
188 : #elif BITSET_WORD_MAX == (0xffffffff + 2) * 0xffffffff
189 : /* Work around a bug in 64-bit PGC (before version 6.1-2), where the
190 : preprocessor mishandles large unsigned values as if they were signed. */
191 : # define BITSET_WORD_BITS 64
192 : #else
193 : # error "Add case for new bitset_word_t size"
194 : #endif
195 :
196 : /* Number of bitset_word_t values in a bitset_t. */
197 : #define BITSET_WORDS ((SBC_MAX + BITSET_WORD_BITS - 1) / BITSET_WORD_BITS)
198 :
199 : typedef bitset_word_t bitset_t[BITSET_WORDS];
200 : typedef bitset_word_t *re_bitset_ptr_t;
201 : typedef const bitset_word_t *re_const_bitset_ptr_t;
202 :
203 : #define PREV_WORD_CONSTRAINT 0x0001
204 : #define PREV_NOTWORD_CONSTRAINT 0x0002
205 : #define NEXT_WORD_CONSTRAINT 0x0004
206 : #define NEXT_NOTWORD_CONSTRAINT 0x0008
207 : #define PREV_NEWLINE_CONSTRAINT 0x0010
208 : #define NEXT_NEWLINE_CONSTRAINT 0x0020
209 : #define PREV_BEGBUF_CONSTRAINT 0x0040
210 : #define NEXT_ENDBUF_CONSTRAINT 0x0080
211 : #define WORD_DELIM_CONSTRAINT 0x0100
212 : #define NOT_WORD_DELIM_CONSTRAINT 0x0200
213 :
214 : typedef enum
215 : {
216 : INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
217 : WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
218 : WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
219 : INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
220 : LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
221 : LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
222 : BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
223 : BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
224 : WORD_DELIM = WORD_DELIM_CONSTRAINT,
225 : NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
226 : } re_context_type;
227 :
228 : typedef struct
229 : {
230 : Idx alloc;
231 : Idx nelem;
232 : Idx *elems;
233 : } re_node_set;
234 :
235 : typedef enum
236 : {
237 : NON_TYPE = 0,
238 :
239 : /* Node type, These are used by token, node, tree. */
240 : CHARACTER = 1,
241 : END_OF_RE = 2,
242 : SIMPLE_BRACKET = 3,
243 : OP_BACK_REF = 4,
244 : OP_PERIOD = 5,
245 : #ifdef RE_ENABLE_I18N
246 : COMPLEX_BRACKET = 6,
247 : OP_UTF8_PERIOD = 7,
248 : #endif /* RE_ENABLE_I18N */
249 :
250 : /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
251 : when the debugger shows values of this enum type. */
252 : #define EPSILON_BIT 8
253 : OP_OPEN_SUBEXP = EPSILON_BIT | 0,
254 : OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
255 : OP_ALT = EPSILON_BIT | 2,
256 : OP_DUP_ASTERISK = EPSILON_BIT | 3,
257 : ANCHOR = EPSILON_BIT | 4,
258 :
259 : /* Tree type, these are used only by tree. */
260 : CONCAT = 16,
261 : SUBEXP = 17,
262 :
263 : /* Token type, these are used only by token. */
264 : OP_DUP_PLUS = 18,
265 : OP_DUP_QUESTION,
266 : OP_OPEN_BRACKET,
267 : OP_CLOSE_BRACKET,
268 : OP_CHARSET_RANGE,
269 : OP_OPEN_DUP_NUM,
270 : OP_CLOSE_DUP_NUM,
271 : OP_NON_MATCH_LIST,
272 : OP_OPEN_COLL_ELEM,
273 : OP_CLOSE_COLL_ELEM,
274 : OP_OPEN_EQUIV_CLASS,
275 : OP_CLOSE_EQUIV_CLASS,
276 : OP_OPEN_CHAR_CLASS,
277 : OP_CLOSE_CHAR_CLASS,
278 : OP_WORD,
279 : OP_NOTWORD,
280 : OP_SPACE,
281 : OP_NOTSPACE,
282 : BACK_SLASH
283 :
284 : } re_token_type_t;
285 :
286 : #ifdef RE_ENABLE_I18N
287 : typedef struct
288 : {
289 : /* Multibyte characters. */
290 : wchar_t *mbchars;
291 :
292 : /* Collating symbols. */
293 : # ifdef _LIBC
294 : int32_t *coll_syms;
295 : # endif
296 :
297 : /* Equivalence classes. */
298 : # ifdef _LIBC
299 : int32_t *equiv_classes;
300 : # endif
301 :
302 : /* Range expressions. */
303 : # ifdef _LIBC
304 : uint32_t *range_starts;
305 : uint32_t *range_ends;
306 : # else /* not _LIBC */
307 : wchar_t *range_starts;
308 : wchar_t *range_ends;
309 : # endif /* not _LIBC */
310 :
311 : /* Character classes. */
312 : wctype_t *char_classes;
313 :
314 : /* If this character set is the non-matching list. */
315 : unsigned int non_match : 1;
316 :
317 : /* # of multibyte characters. */
318 : Idx nmbchars;
319 :
320 : /* # of collating symbols. */
321 : Idx ncoll_syms;
322 :
323 : /* # of equivalence classes. */
324 : Idx nequiv_classes;
325 :
326 : /* # of range expressions. */
327 : Idx nranges;
328 :
329 : /* # of character classes. */
330 : Idx nchar_classes;
331 : } re_charset_t;
332 : #endif /* RE_ENABLE_I18N */
333 :
334 : typedef struct
335 : {
336 : union
337 : {
338 : unsigned char c; /* for CHARACTER */
339 : re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
340 : #ifdef RE_ENABLE_I18N
341 : re_charset_t *mbcset; /* for COMPLEX_BRACKET */
342 : #endif /* RE_ENABLE_I18N */
343 : Idx idx; /* for BACK_REF */
344 : re_context_type ctx_type; /* for ANCHOR */
345 : } opr;
346 : #if __GNUC__ >= 2 && !__STRICT_ANSI__
347 : re_token_type_t type : 8;
348 : #else
349 : re_token_type_t type;
350 : #endif
351 : unsigned int constraint : 10; /* context constraint */
352 : unsigned int duplicated : 1;
353 : unsigned int opt_subexp : 1;
354 : #ifdef RE_ENABLE_I18N
355 : unsigned int accept_mb : 1;
356 : /* These 2 bits can be moved into the union if needed (e.g. if running out
357 : of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
358 : unsigned int mb_partial : 1;
359 : #endif
360 : unsigned int word_char : 1;
361 : } re_token_t;
362 :
363 : #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
364 :
365 : struct re_string_t
366 : {
367 : /* Indicate the raw buffer which is the original string passed as an
368 : argument of regexec(), re_search(), etc.. */
369 : const unsigned char *raw_mbs;
370 : /* Store the multibyte string. In case of "case insensitive mode" like
371 : REG_ICASE, upper cases of the string are stored, otherwise MBS points
372 : the same address that RAW_MBS points. */
373 : unsigned char *mbs;
374 : #ifdef RE_ENABLE_I18N
375 : /* Store the wide character string which is corresponding to MBS. */
376 : wint_t *wcs;
377 : Idx *offsets;
378 : mbstate_t cur_state;
379 : #endif
380 : /* Index in RAW_MBS. Each character mbs[i] corresponds to
381 : raw_mbs[raw_mbs_idx + i]. */
382 : Idx raw_mbs_idx;
383 : /* The length of the valid characters in the buffers. */
384 : Idx valid_len;
385 : /* The corresponding number of bytes in raw_mbs array. */
386 : Idx valid_raw_len;
387 : /* The length of the buffers MBS and WCS. */
388 : Idx bufs_len;
389 : /* The index in MBS, which is updated by re_string_fetch_byte. */
390 : Idx cur_idx;
391 : /* length of RAW_MBS array. */
392 : Idx raw_len;
393 : /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
394 : Idx len;
395 : /* End of the buffer may be shorter than its length in the cases such
396 : as re_match_2, re_search_2. Then, we use STOP for end of the buffer
397 : instead of LEN. */
398 : Idx raw_stop;
399 : /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
400 : Idx stop;
401 :
402 : /* The context of mbs[0]. We store the context independently, since
403 : the context of mbs[0] may be different from raw_mbs[0], which is
404 : the beginning of the input string. */
405 : unsigned int tip_context;
406 : /* The translation passed as a part of an argument of re_compile_pattern. */
407 : RE_TRANSLATE_TYPE trans;
408 : /* Copy of re_dfa_t's word_char. */
409 : re_const_bitset_ptr_t word_char;
410 : /* true if REG_ICASE. */
411 : unsigned char icase;
412 : unsigned char is_utf8;
413 : unsigned char map_notascii;
414 : unsigned char mbs_allocated;
415 : unsigned char offsets_needed;
416 : unsigned char newline_anchor;
417 : unsigned char word_ops_used;
418 : int mb_cur_max;
419 : };
420 : typedef struct re_string_t re_string_t;
421 :
422 :
423 : struct re_dfa_t;
424 : typedef struct re_dfa_t re_dfa_t;
425 :
426 : #ifndef _LIBC
427 : # if defined __i386__ && !defined __EMX__
428 : # define internal_function __attribute ((regparm (3), stdcall))
429 : # else
430 : # define internal_function
431 : # endif
432 : #endif
433 :
434 : static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
435 : Idx new_buf_len)
436 : internal_function;
437 : #ifdef RE_ENABLE_I18N
438 : static void build_wcs_buffer (re_string_t *pstr) internal_function;
439 : static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr)
440 : internal_function;
441 : #endif /* RE_ENABLE_I18N */
442 : static void build_upper_buffer (re_string_t *pstr) internal_function;
443 : static void re_string_translate_buffer (re_string_t *pstr) internal_function;
444 : static unsigned int re_string_context_at (const re_string_t *input, Idx idx,
445 : int eflags)
446 : internal_function __attribute ((pure));
447 : #define re_string_peek_byte(pstr, offset) \
448 : ((pstr)->mbs[(pstr)->cur_idx + offset])
449 : #define re_string_fetch_byte(pstr) \
450 : ((pstr)->mbs[(pstr)->cur_idx++])
451 : #define re_string_first_byte(pstr, idx) \
452 : ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
453 : #define re_string_is_single_byte_char(pstr, idx) \
454 : ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
455 : || (pstr)->wcs[(idx) + 1] != WEOF))
456 : #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
457 : #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
458 : #define re_string_get_buffer(pstr) ((pstr)->mbs)
459 : #define re_string_length(pstr) ((pstr)->len)
460 : #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
461 : #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
462 : #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
463 :
464 : #include <alloca.h>
465 :
466 : #ifndef _LIBC
467 : # if HAVE_ALLOCA
468 : /* The OS usually guarantees only one guard page at the bottom of the stack,
469 : and a page size can be as small as 4096 bytes. So we cannot safely
470 : allocate anything larger than 4096 bytes. Also care for the possibility
471 : of a few compiler-allocated temporary stack slots. */
472 : # define __libc_use_alloca(n) ((n) < 4032)
473 : # else
474 : /* alloca is implemented with malloc, so just use malloc. */
475 : # define __libc_use_alloca(n) 0
476 : # endif
477 : #endif
478 :
479 : #ifndef MAX
480 : # define MAX(a,b) ((a) < (b) ? (b) : (a))
481 : #endif
482 :
483 : #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
484 : #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
485 : #define re_free(p) free (p)
486 :
487 : struct bin_tree_t
488 : {
489 : struct bin_tree_t *parent;
490 : struct bin_tree_t *left;
491 : struct bin_tree_t *right;
492 : struct bin_tree_t *first;
493 : struct bin_tree_t *next;
494 :
495 : re_token_t token;
496 :
497 : /* `node_idx' is the index in dfa->nodes, if `type' == 0.
498 : Otherwise `type' indicate the type of this node. */
499 : Idx node_idx;
500 : };
501 : typedef struct bin_tree_t bin_tree_t;
502 :
503 : #define BIN_TREE_STORAGE_SIZE \
504 : ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
505 :
506 : struct bin_tree_storage_t
507 : {
508 : struct bin_tree_storage_t *next;
509 : bin_tree_t data[BIN_TREE_STORAGE_SIZE];
510 : };
511 : typedef struct bin_tree_storage_t bin_tree_storage_t;
512 :
513 : #define CONTEXT_WORD 1
514 : #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
515 : #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
516 : #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
517 :
518 : #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
519 : #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
520 : #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
521 : #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
522 : #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
523 :
524 : #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
525 : #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
526 : #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
527 : #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
528 :
529 : #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
530 : ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
531 : || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
532 : || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
533 : || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
534 :
535 : #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
536 : ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
537 : || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
538 : || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
539 : || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
540 :
541 : struct re_dfastate_t
542 : {
543 : re_hashval_t hash;
544 : re_node_set nodes;
545 : re_node_set non_eps_nodes;
546 : re_node_set inveclosure;
547 : re_node_set *entrance_nodes;
548 : struct re_dfastate_t **trtable, **word_trtable;
549 : unsigned int context : 4;
550 : unsigned int halt : 1;
551 : /* If this state can accept `multi byte'.
552 : Note that we refer to multibyte characters, and multi character
553 : collating elements as `multi byte'. */
554 : unsigned int accept_mb : 1;
555 : /* If this state has backreference node(s). */
556 : unsigned int has_backref : 1;
557 : unsigned int has_constraint : 1;
558 : };
559 : typedef struct re_dfastate_t re_dfastate_t;
560 :
561 : struct re_state_table_entry
562 : {
563 : Idx num;
564 : Idx alloc;
565 : re_dfastate_t **array;
566 : };
567 :
568 : /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
569 :
570 : typedef struct
571 : {
572 : Idx next_idx;
573 : Idx alloc;
574 : re_dfastate_t **array;
575 : } state_array_t;
576 :
577 : /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
578 :
579 : typedef struct
580 : {
581 : Idx node;
582 : Idx str_idx; /* The position NODE match at. */
583 : state_array_t path;
584 : } re_sub_match_last_t;
585 :
586 : /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
587 : And information about the node, whose type is OP_CLOSE_SUBEXP,
588 : corresponding to NODE is stored in LASTS. */
589 :
590 : typedef struct
591 : {
592 : Idx str_idx;
593 : Idx node;
594 : state_array_t *path;
595 : Idx alasts; /* Allocation size of LASTS. */
596 : Idx nlasts; /* The number of LASTS. */
597 : re_sub_match_last_t **lasts;
598 : } re_sub_match_top_t;
599 :
600 : struct re_backref_cache_entry
601 : {
602 : Idx node;
603 : Idx str_idx;
604 : Idx subexp_from;
605 : Idx subexp_to;
606 : char more;
607 : char unused;
608 : unsigned short int eps_reachable_subexps_map;
609 : };
610 :
611 : typedef struct
612 : {
613 : /* The string object corresponding to the input string. */
614 : re_string_t input;
615 : #if defined _LIBC || (defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L)
616 : const re_dfa_t *const dfa;
617 : #else
618 : const re_dfa_t *dfa;
619 : #endif
620 : /* EFLAGS of the argument of regexec. */
621 : int eflags;
622 : /* Where the matching ends. */
623 : Idx match_last;
624 : Idx last_node;
625 : /* The state log used by the matcher. */
626 : re_dfastate_t **state_log;
627 : Idx state_log_top;
628 : /* Back reference cache. */
629 : Idx nbkref_ents;
630 : Idx abkref_ents;
631 : struct re_backref_cache_entry *bkref_ents;
632 : int max_mb_elem_len;
633 : Idx nsub_tops;
634 : Idx asub_tops;
635 : re_sub_match_top_t **sub_tops;
636 : } re_match_context_t;
637 :
638 : typedef struct
639 : {
640 : re_dfastate_t **sifted_states;
641 : re_dfastate_t **limited_states;
642 : Idx last_node;
643 : Idx last_str_idx;
644 : re_node_set limits;
645 : } re_sift_context_t;
646 :
647 : struct re_fail_stack_ent_t
648 : {
649 : Idx idx;
650 : Idx node;
651 : regmatch_t *regs;
652 : re_node_set eps_via_nodes;
653 : };
654 :
655 : struct re_fail_stack_t
656 : {
657 : Idx num;
658 : Idx alloc;
659 : struct re_fail_stack_ent_t *stack;
660 : };
661 :
662 : struct re_dfa_t
663 : {
664 : re_token_t *nodes;
665 : size_t nodes_alloc;
666 : size_t nodes_len;
667 : Idx *nexts;
668 : Idx *org_indices;
669 : re_node_set *edests;
670 : re_node_set *eclosures;
671 : re_node_set *inveclosures;
672 : struct re_state_table_entry *state_table;
673 : re_dfastate_t *init_state;
674 : re_dfastate_t *init_state_word;
675 : re_dfastate_t *init_state_nl;
676 : re_dfastate_t *init_state_begbuf;
677 : bin_tree_t *str_tree;
678 : bin_tree_storage_t *str_tree_storage;
679 : re_bitset_ptr_t sb_char;
680 : int str_tree_storage_idx;
681 :
682 : /* number of subexpressions `re_nsub' is in regex_t. */
683 : re_hashval_t state_hash_mask;
684 : Idx init_node;
685 : Idx nbackref; /* The number of backreference in this dfa. */
686 :
687 : /* Bitmap expressing which backreference is used. */
688 : bitset_word_t used_bkref_map;
689 : bitset_word_t completed_bkref_map;
690 :
691 : unsigned int has_plural_match : 1;
692 : /* If this dfa has "multibyte node", which is a backreference or
693 : a node which can accept multibyte character or multi character
694 : collating element. */
695 : unsigned int has_mb_node : 1;
696 : unsigned int is_utf8 : 1;
697 : unsigned int map_notascii : 1;
698 : unsigned int word_ops_used : 1;
699 : int mb_cur_max;
700 : bitset_t word_char;
701 : reg_syntax_t syntax;
702 : Idx *subexp_map;
703 : #ifdef DEBUG
704 : char* re_str;
705 : #endif
706 : #ifdef _LIBC
707 : __libc_lock_define (, lock)
708 : #endif
709 : };
710 :
711 : #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
712 : #define re_node_set_remove(set,id) \
713 : (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
714 : #define re_node_set_empty(p) ((p)->nelem = 0)
715 : #define re_node_set_free(set) re_free ((set)->elems)
716 :
717 :
718 : typedef enum
719 : {
720 : SB_CHAR,
721 : MB_CHAR,
722 : EQUIV_CLASS,
723 : COLL_SYM,
724 : CHAR_CLASS
725 : } bracket_elem_type;
726 :
727 : typedef struct
728 : {
729 : bracket_elem_type type;
730 : union
731 : {
732 : unsigned char ch;
733 : unsigned char *name;
734 : wchar_t wch;
735 : } opr;
736 : } bracket_elem_t;
737 :
738 :
739 1605 : /* Inline functions for bitset_t operation. */
740 :
741 1605 : static inline void
742 1605 : bitset_set (bitset_t set, Idx i)
743 : {
744 : set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS;
745 2 : }
746 :
747 2 : static inline void
748 2 : bitset_clear (bitset_t set, Idx i)
749 : {
750 : set[i / BITSET_WORD_BITS] &= ~ ((bitset_word_t) 1 << i % BITSET_WORD_BITS);
751 388 : }
752 :
753 388 : static inline bool
754 : bitset_contain (const bitset_t set, Idx i)
755 : {
756 : return (set[i / BITSET_WORD_BITS] >> i % BITSET_WORD_BITS) & 1;
757 307 : }
758 :
759 307 : static inline void
760 307 : bitset_empty (bitset_t set)
761 : {
762 : memset (set, '\0', sizeof (bitset_t));
763 2 : }
764 :
765 2 : static inline void
766 : bitset_set_all (bitset_t set)
767 : {
768 : memset (set, -1, sizeof (bitset_word_t) * (SBC_MAX / BITSET_WORD_BITS));
769 2 : if (SBC_MAX % BITSET_WORD_BITS != 0)
770 : set[BITSET_WORDS - 1] =
771 : ((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1;
772 93 : }
773 :
774 93 : static inline void
775 93 : bitset_copy (bitset_t dest, const bitset_t src)
776 : {
777 : memcpy (dest, src, sizeof (bitset_t));
778 13 : }
779 :
780 : static inline void
781 65 : bitset_not (bitset_t set)
782 52 : {
783 : int bitset_i;
784 : for (bitset_i = 0; bitset_i < SBC_MAX / BITSET_WORD_BITS; ++bitset_i)
785 : set[bitset_i] = ~set[bitset_i];
786 : if (SBC_MAX % BITSET_WORD_BITS != 0)
787 13 : set[BITSET_WORDS - 1] =
788 : ((((bitset_word_t) 1 << SBC_MAX % BITSET_WORD_BITS) - 1)
789 : & ~set[BITSET_WORDS - 1]);
790 128 : }
791 :
792 : static inline void
793 640 : bitset_merge (bitset_t dest, const bitset_t src)
794 512 : {
795 128 : int bitset_i;
796 : for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
797 : dest[bitset_i] |= src[bitset_i];
798 0 : }
799 :
800 : static inline void
801 0 : bitset_mask (bitset_t dest, const bitset_t src)
802 0 : {
803 0 : int bitset_i;
804 : for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
805 : dest[bitset_i] &= src[bitset_i];
806 : }
807 :
808 : #ifdef RE_ENABLE_I18N
809 814 : /* Inline functions for re_string. */
810 : static inline int
811 : internal_function __attribute ((pure))
812 814 : re_string_char_size_at (const re_string_t *pstr, Idx idx)
813 814 : {
814 0 : int byte_idx;
815 0 : if (pstr->mb_cur_max == 1)
816 0 : return 1;
817 0 : for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
818 : if (pstr->wcs[idx + byte_idx] != WEOF)
819 : break;
820 : return byte_idx;
821 : }
822 0 :
823 : static inline wint_t
824 0 : internal_function __attribute ((pure))
825 0 : re_string_wchar_at (const re_string_t *pstr, Idx idx)
826 0 : {
827 : if (pstr->mb_cur_max == 1)
828 : return (wint_t) pstr->mbs[idx];
829 : return (wint_t) pstr->wcs[idx];
830 : }
831 0 :
832 : static int
833 : internal_function __attribute ((pure))
834 : re_string_elem_size_at (const re_string_t *pstr, Idx idx)
835 : {
836 : # ifdef _LIBC
837 : const unsigned char *p, *extra;
838 : const int32_t *table, *indirect;
839 : int32_t tmp;
840 : # include <locale/weight.h>
841 : uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
842 :
843 : if (nrules != 0)
844 : {
845 : table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
846 : extra = (const unsigned char *)
847 : _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
848 : indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
849 : _NL_COLLATE_INDIRECTMB);
850 : p = pstr->mbs + idx;
851 : tmp = findidx (&p);
852 : return p - pstr->mbs - idx;
853 0 : }
854 : else
855 : # endif /* _LIBC */
856 : return 1;
857 : }
858 : #endif /* RE_ENABLE_I18N */
859 :
860 : #endif /* _REGEX_INTERNAL_H */
|
