1 /* Extended regular expression matching and search library.
2 Copyright (C) 2002 Free Software Foundation, Inc.
3 This file is part of the GNU C Library.
4 Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
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.
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.
16 You should have received a copy of the GNU Lesser General Public
17 License along with the GNU C Library; if not, write to the Free
18 Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
21 #ifndef _REGEX_INTERNAL_H
22 #define _REGEX_INTERNAL_H 1
24 /* Number of bits in a byte. */
26 /* Number of single byte character. */
29 #define COLL_ELEM_LEN_MAX 8
31 /* The character which represents newline. */
32 #define NEWLINE_CHAR '\n'
34 /* Rename to standard API for using out of glibc. */
36 # define __wctype wctype
37 # define __iswctype iswctype
38 # define __btowc btowc
39 # define __mempcpy memcpy
40 # define attribute_hidden
41 #endif /* not _LIBC */
43 extern const char __re_error_msgid
[] attribute_hidden
;
44 extern const size_t __re_error_msgid_idx
[] attribute_hidden
;
46 /* Number of bits in an unsinged int. */
47 #define UINT_BITS (sizeof (unsigned int) * BYTE_BITS)
48 /* Number of unsigned int in an bit_set. */
49 #define BITSET_UINTS ((SBC_MAX + UINT_BITS - 1) / UINT_BITS)
50 typedef unsigned int bitset
[BITSET_UINTS
];
51 typedef unsigned int *re_bitset_ptr_t
;
53 #define bitset_set(set,i) (set[i / UINT_BITS] |= 1 << i % UINT_BITS)
54 #define bitset_clear(set,i) (set[i / UINT_BITS] &= ~(1 << i % UINT_BITS))
55 #define bitset_contain(set,i) (set[i / UINT_BITS] & (1 << i % UINT_BITS))
56 #define bitset_empty(set) memset (set, 0, sizeof (unsigned int) * BITSET_UINTS)
57 #define bitset_set_all(set) \
58 memset (set, 255, sizeof (unsigned int) * BITSET_UINTS)
59 #define bitset_copy(dest,src) \
60 memcpy (dest, src, sizeof (unsigned int) * BITSET_UINTS)
61 static inline void bitset_not (bitset set
);
62 static inline void bitset_merge (bitset dest
, const bitset src
);
63 static inline void bitset_not_merge (bitset dest
, const bitset src
);
65 #define PREV_WORD_CONSTRAINT 0x0001
66 #define PREV_NOTWORD_CONSTRAINT 0x0002
67 #define NEXT_WORD_CONSTRAINT 0x0004
68 #define NEXT_NOTWORD_CONSTRAINT 0x0008
69 #define PREV_NEWLINE_CONSTRAINT 0x0010
70 #define NEXT_NEWLINE_CONSTRAINT 0x0020
71 #define PREV_BEGBUF_CONSTRAINT 0x0040
72 #define NEXT_ENDBUF_CONSTRAINT 0x0080
73 #define DUMMY_CONSTRAINT 0x0100
77 INSIDE_WORD
= PREV_WORD_CONSTRAINT
| NEXT_WORD_CONSTRAINT
,
78 WORD_FIRST
= PREV_NOTWORD_CONSTRAINT
| NEXT_WORD_CONSTRAINT
,
79 WORD_LAST
= PREV_WORD_CONSTRAINT
| NEXT_NOTWORD_CONSTRAINT
,
80 LINE_FIRST
= PREV_NEWLINE_CONSTRAINT
,
81 LINE_LAST
= NEXT_NEWLINE_CONSTRAINT
,
82 BUF_FIRST
= PREV_BEGBUF_CONSTRAINT
,
83 BUF_LAST
= NEXT_ENDBUF_CONSTRAINT
,
84 WORD_DELIM
= DUMMY_CONSTRAINT
98 /* Token type, these are used only by token. */
110 OP_CLOSE_EQUIV_CLASS
,
117 /* Tree type, these are used only by tree. */
124 /* Node type, These are used by token, node, tree. */
142 /* If this character set is the non-matching list. */
143 unsigned int non_match
: 1;
145 /* Multibyte characters. */
149 /* Collating symbols. */
155 /* Equivalence classes. */
157 int32_t *equiv_classes
;
161 /* Range expressions. */
163 uint32_t *range_starts
;
164 uint32_t *range_ends
;
165 #else /* not _LIBC */
166 wchar_t *range_starts
;
168 #endif /* not _LIBC */
171 /* Character classes. */
172 wctype_t *char_classes
;
178 re_token_type_t type
;
181 unsigned char c
; /* for CHARACTER */
182 re_bitset_ptr_t sbcset
; /* for SIMPLE_BRACKET */
183 re_charset_t
*mbcset
; /* for COMPLEX_BRACKET */
184 int idx
; /* for BACK_REF */
185 re_context_type ctx_type
; /* for ANCHOR */
188 int entity
; /* for OP_CONTEXT_NODE, index of the entity */
189 re_node_set
*bkref_eclosure
;
192 unsigned int constraint
: 10; /* context constraint */
193 unsigned int duplicated
: 1;
194 #ifdef RE_ENABLE_I18N
195 unsigned int mb_partial
: 1;
199 #define IS_EPSILON_NODE(type) \
200 ((type) == OP_ALT || (type) == OP_DUP_ASTERISK || (type) == OP_DUP_PLUS || \
201 (type) == OP_DUP_QUESTION || (type) == ANCHOR)
203 #define ACCEPT_MB_NODE(type) \
204 ((type) == COMPLEX_BRACKET || (type) == OP_PERIOD)
208 /* Indicate the raw buffer which is the original string passed as an
209 argument of regexec(), re_search(), etc.. */
210 const unsigned char *raw_mbs
;
211 /* Index in RAW_MBS. Each character mbs[i] corresponds to
212 raw_mbs[raw_mbs_idx + i]. */
214 /* Store the multibyte string. In case of "case insensitive mode" like
215 REG_ICASE, upper cases of the string are stored, otherwise MBS points
216 the same address that RAW_MBS points. */
218 /* Store the case sensitive multibyte string. In case of
219 "case insensitive mode", the original string are stored,
220 otherwise MBS_CASE points the same address that MBS points. */
221 unsigned char *mbs_case
;
222 #ifdef RE_ENABLE_I18N
223 /* Store the wide character string which is corresponding to MBS. */
227 /* The length of the valid characters in the buffers. */
229 /* The length of the buffers MBS, MBS_CASE, and WCS. */
231 /* The index in MBS, which is updated by re_string_fetch_byte. */
233 /* This is length_of_RAW_MBS - RAW_MBS_IDX. */
235 /* The context of mbs[0]. We store the context independently, since
236 the context of mbs[0] may be different from raw_mbs[0], which is
237 the beginning of the input string. */
238 unsigned int tip_context
;
239 /* The translation passed as a part of an argument of re_compile_pattern. */
240 RE_TRANSLATE_TYPE trans
;
241 /* 1 if REG_ICASE. */
242 unsigned int icase
: 1;
244 typedef struct re_string_t re_string_t
;
245 /* In case of REG_ICASE, we allocate the buffer dynamically for mbs. */
246 #define MBS_ALLOCATED(pstr) (pstr->icase)
247 /* In case that we need translation, we allocate the buffer dynamically
248 for mbs_case. Note that mbs == mbs_case if not REG_ICASE. */
249 #define MBS_CASE_ALLOCATED(pstr) (pstr->trans != NULL)
252 static reg_errcode_t
re_string_allocate (re_string_t
*pstr
,
253 const unsigned char *str
, int len
,
255 RE_TRANSLATE_TYPE trans
, int icase
);
256 static reg_errcode_t
re_string_construct (re_string_t
*pstr
,
257 const unsigned char *str
, int len
,
258 RE_TRANSLATE_TYPE trans
, int icase
);
259 static reg_errcode_t
re_string_reconstruct (re_string_t
*pstr
, int idx
,
260 int eflags
, int newline
);
261 static reg_errcode_t
re_string_realloc_buffers (re_string_t
*pstr
,
263 #ifdef RE_ENABLE_I18N
264 static void build_wcs_buffer (re_string_t
*pstr
);
265 static void build_wcs_upper_buffer (re_string_t
*pstr
);
266 #endif /* RE_ENABLE_I18N */
267 static void build_upper_buffer (re_string_t
*pstr
);
268 static void re_string_translate_buffer (re_string_t
*pstr
);
269 static void re_string_destruct (re_string_t
*pstr
);
270 #ifdef RE_ENABLE_I18N
271 static int re_string_elem_size_at (const re_string_t
*pstr
, int idx
);
272 static inline int re_string_char_size_at (const re_string_t
*pstr
, int idx
);
273 static inline wint_t re_string_wchar_at (const re_string_t
*pstr
, int idx
);
274 #endif /* RE_ENABLE_I18N */
275 static unsigned int re_string_context_at (const re_string_t
*input
, int idx
,
276 int eflags
, int newline_anchor
);
277 #define re_string_peek_byte(pstr, offset) \
278 ((pstr)->mbs[(pstr)->cur_idx + offset])
279 #define re_string_peek_byte_case(pstr, offset) \
280 ((pstr)->mbs_case[(pstr)->cur_idx + offset])
281 #define re_string_fetch_byte(pstr) \
282 ((pstr)->mbs[(pstr)->cur_idx++])
283 #define re_string_fetch_byte_case(pstr) \
284 ((pstr)->mbs_case[(pstr)->cur_idx++])
285 #define re_string_first_byte(pstr, idx) \
286 ((idx) == (pstr)->len || (pstr)->wcs[idx] != WEOF)
287 #define re_string_is_single_byte_char(pstr, idx) \
288 ((pstr)->wcs[idx] != WEOF && ((pstr)->len == (idx) \
289 || (pstr)->wcs[(idx) + 1] != WEOF))
290 #define re_string_eoi(pstr) ((pstr)->len == (pstr)->cur_idx)
291 #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
292 #define re_string_get_buffer(pstr) ((pstr)->mbs)
293 #define re_string_length(pstr) ((pstr)->len)
294 #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
295 #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
296 #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
298 #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
299 #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
300 #define re_free(p) free (p)
304 struct bin_tree_t
*parent
;
305 struct bin_tree_t
*left
;
306 struct bin_tree_t
*right
;
308 /* `node_idx' is the index in dfa->nodes, if `type' == 0.
309 Otherwise `type' indicate the type of this node. */
310 re_token_type_t type
;
315 re_node_set eclosure
;
317 typedef struct bin_tree_t bin_tree_t
;
320 #define CONTEXT_WORD 1
321 #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
322 #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
323 #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
325 #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
326 #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
327 #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
328 #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
329 #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
331 #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
332 #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
334 #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
335 ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
336 || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
337 || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
338 || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
340 #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
341 ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
342 || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
343 || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
344 || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
350 re_node_set
*entrance_nodes
;
351 struct re_dfastate_t
**trtable
;
352 struct re_dfastate_t
**trtable_search
;
353 /* If this state is a special state.
354 A state is a special state if the state is the halt state, or
356 unsigned int context
: 2;
357 unsigned int halt
: 1;
358 /* If this state can accept `multi byte'.
359 Note that we refer to multibyte characters, and multi character
360 collating elements as `multi byte'. */
361 unsigned int accept_mb
: 1;
362 /* If this state has backreference node(s). */
363 unsigned int has_backref
: 1;
364 unsigned int has_constraint
: 1;
366 typedef struct re_dfastate_t re_dfastate_t
;
370 /* start <= node < end */
375 struct re_state_table_entry
379 re_dfastate_t
**array
;
382 struct re_backref_cache_entry
392 /* EFLAGS of the argument of regexec. */
394 /* Where the matching ends. */
396 /* The string object corresponding to the input string. */
398 /* The state log used by the matcher. */
399 re_dfastate_t
**state_log
;
401 /* Back reference cache. */
404 struct re_backref_cache_entry
*bkref_ents
;
406 } re_match_context_t
;
410 re_bitset_ptr_t word_char
;
412 /* number of subexpressions `re_nsub' is in regex_t. */
414 re_subexp_t
*subexps
;
419 bin_tree_t
*str_tree
;
423 re_node_set
*eclosures
;
424 re_node_set
*inveclosures
;
425 struct re_state_table_entry
*state_table
;
426 unsigned int state_hash_mask
;
427 re_dfastate_t
*init_state
;
428 re_dfastate_t
*init_state_word
;
429 re_dfastate_t
*init_state_nl
;
430 re_dfastate_t
*init_state_begbuf
;
433 int nbackref
; /* The number of backreference in this dfa. */
434 /* If this dfa has "multibyte node", which is a backreference or
435 a node which can accept multibyte character or multi character
436 collating element. */
437 unsigned int has_mb_node
: 1;
439 typedef struct re_dfa_t re_dfa_t
;
441 static reg_errcode_t
re_node_set_alloc (re_node_set
*set
, int size
);
442 static reg_errcode_t
re_node_set_init_1 (re_node_set
*set
, int elem
);
443 static reg_errcode_t
re_node_set_init_2 (re_node_set
*set
, int elem1
,
445 #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
446 static reg_errcode_t
re_node_set_init_copy (re_node_set
*dest
,
447 const re_node_set
*src
);
448 static reg_errcode_t
re_node_set_intersect (re_node_set
*dest
,
449 const re_node_set
*src1
,
450 const re_node_set
*src2
);
451 static reg_errcode_t
re_node_set_add_intersect (re_node_set
*dest
,
452 const re_node_set
*src1
,
453 const re_node_set
*src2
);
454 static reg_errcode_t
re_node_set_init_union (re_node_set
*dest
,
455 const re_node_set
*src1
,
456 const re_node_set
*src2
);
457 static reg_errcode_t
re_node_set_merge (re_node_set
*dest
,
458 const re_node_set
*src
);
459 static int re_node_set_insert (re_node_set
*set
, int elem
);
460 static int re_node_set_compare (const re_node_set
*set1
,
461 const re_node_set
*set2
);
462 static int re_node_set_contains (const re_node_set
*set
, int elem
);
463 static void re_node_set_remove_at (re_node_set
*set
, int idx
);
464 #define re_node_set_empty(p) ((p)->nelem = 0)
465 #define re_node_set_free(set) re_free ((set)->elems)
466 static int re_dfa_add_node (re_dfa_t
*dfa
, re_token_t token
, int mode
);
467 static re_dfastate_t
*re_acquire_state (reg_errcode_t
*err
, re_dfa_t
*dfa
,
468 const re_node_set
*nodes
);
469 static re_dfastate_t
*re_acquire_state_context (reg_errcode_t
*err
,
471 const re_node_set
*nodes
,
472 unsigned int context
);
486 bracket_elem_type type
;
496 /* Inline functions for bitset operation. */
502 for (bitset_i
= 0; bitset_i
< BITSET_UINTS
; ++bitset_i
)
503 set
[bitset_i
] = ~set
[bitset_i
];
507 bitset_merge (dest
, src
)
512 for (bitset_i
= 0; bitset_i
< BITSET_UINTS
; ++bitset_i
)
513 dest
[bitset_i
] |= src
[bitset_i
];
517 bitset_not_merge (dest
, src
)
522 for (i
= 0; i
< BITSET_UINTS
; ++i
)
526 #ifdef RE_ENABLE_I18N
527 /* Inline functions for re_string. */
529 re_string_char_size_at (pstr
, idx
)
530 const re_string_t
*pstr
;
536 for (byte_idx
= 1; idx
+ byte_idx
< pstr
->len
; ++byte_idx
)
537 if (pstr
->wcs
[idx
+ byte_idx
] != WEOF
)
543 re_string_wchar_at (pstr
, idx
)
544 const re_string_t
*pstr
;
548 return (wint_t) pstr
->mbs
[idx
];
549 return (wint_t) pstr
->wcs
[idx
];
553 re_string_elem_size_at (pstr
, idx
)
554 const re_string_t
*pstr
;
558 const unsigned char *p
;
560 const int32_t *table
, *indirect
;
562 # include <locale/weight.h>
563 uint_fast32_t nrules
= _NL_CURRENT_WORD (LC_COLLATE
, _NL_COLLATE_NRULES
);
567 table
= (const int32_t *) _NL_CURRENT (LC_COLLATE
, _NL_COLLATE_TABLEMB
);
568 extra
= (const char *) _NL_CURRENT (LC_COLLATE
, _NL_COLLATE_EXTRAMB
);
569 indirect
= (const int32_t *) _NL_CURRENT (LC_COLLATE
,
570 _NL_COLLATE_INDIRECTMB
);
573 return p
- (const unsigned char *) pstr
->mbs
- idx
;
579 #endif /* RE_ENABLE_I18N */
581 #endif /* _REGEX_INTERNAL_H */