regex_internal.h 19 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704
  1. /* Extended regular expression matching and search library.
  2. Copyright (C) 2002, 2003, 2004, 2005 Free Software Foundation, Inc.
  3. This file is part of the GNU C Library.
  4. Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.
  5. The GNU C Library is free software; you can redistribute it and/or
  6. modify it under the terms of the GNU Lesser General Public
  7. License as published by the Free Software Foundation; either
  8. version 2.1 of the License, or (at your option) any later version.
  9. The GNU C Library is distributed in the hope that it will be useful,
  10. but WITHOUT ANY WARRANTY; without even the implied warranty of
  11. MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  12. Lesser General Public License for more details.
  13. You should have received a copy of the GNU Lesser General Public
  14. License along with the GNU C Library; if not, write to the Free
  15. Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  16. 02111-1307 USA. */
  17. #ifndef _REGEX_INTERNAL_H
  18. #define _REGEX_INTERNAL_H 1
  19. #include <assert.h>
  20. #include <ctype.h>
  21. #include <stdio.h>
  22. #include <stdlib.h>
  23. #include <string.h>
  24. #if defined HAVE_LANGINFO_H || defined HAVE_LANGINFO_CODESET
  25. # include <langinfo.h>
  26. #endif
  27. #if defined HAVE_LOCALE_H
  28. # include <locale.h>
  29. #endif
  30. #if defined HAVE_WCHAR_H
  31. # include <wchar.h>
  32. #endif
  33. #if defined HAVE_WCTYPE_H
  34. # include <wctype.h>
  35. #endif
  36. #if defined HAVE_STDBOOL_H
  37. # include <stdbool.h>
  38. #endif
  39. #if defined HAVE_STDINT_H
  40. # include <stdint.h>
  41. #endif
  42. #define __libc_lock_define(CLASS, NAME)
  43. #define __libc_lock_init(NAME) do { } while (0)
  44. #define __libc_lock_lock(NAME) do { } while (0)
  45. #define __libc_lock_unlock(NAME) do { } while (0)
  46. #undef gettext
  47. #undef gettext_noop
  48. #define gettext(msgid) (msgid)
  49. #define gettext_noop(String) String
  50. #if (defined MB_CUR_MAX && HAVE_LOCALE_H && HAVE_WCTYPE_H && HAVE_WCHAR_H && HAVE_WCRTOMB && HAVE_MBRTOWC && HAVE_WCSCOLL)
  51. # define RE_ENABLE_I18N
  52. #endif
  53. #if __GNUC__ >= 3
  54. /* uclibc: lean towards smaller size a bit:
  55. * OFF: # define BE(expr, val) __builtin_expect (expr, val) */
  56. # define BE(expr, val) (expr)
  57. #else
  58. # define BE(expr, val) (expr)
  59. # define inline
  60. #endif
  61. /* Number of single byte character. */
  62. #define SBC_MAX 256
  63. #define COLL_ELEM_LEN_MAX 8
  64. /* The character which represents newline. */
  65. #define NEWLINE_CHAR '\n'
  66. #define WIDE_NEWLINE_CHAR L'\n'
  67. #ifdef __GNUC__
  68. # define __attribute(arg) __attribute__ (arg)
  69. #else
  70. # define __attribute(arg)
  71. #endif
  72. /* An integer used to represent a set of bits. It must be unsigned,
  73. and must be at least as wide as unsigned int. */
  74. typedef unsigned long int bitset_word_t;
  75. /* All bits set in a bitset_word_t. */
  76. #define BITSET_WORD_MAX ULONG_MAX
  77. /* Number of bits in a bitset_word_t. */
  78. #define BITSET_WORD_BITS (sizeof (bitset_word_t) * CHAR_BIT)
  79. /* Number of bitset_word_t in a bit_set. */
  80. #define BITSET_WORDS (SBC_MAX / BITSET_WORD_BITS)
  81. typedef bitset_word_t bitset_t[BITSET_WORDS];
  82. typedef bitset_word_t *re_bitset_ptr_t;
  83. typedef const bitset_word_t *re_const_bitset_ptr_t;
  84. #define bitset_set(set,i) \
  85. (set[i / BITSET_WORD_BITS] |= (bitset_word_t) 1 << i % BITSET_WORD_BITS)
  86. #define bitset_clear(set,i) \
  87. (set[i / BITSET_WORD_BITS] &= ~((bitset_word_t) 1 << i % BITSET_WORD_BITS))
  88. #define bitset_contain(set,i) \
  89. (set[i / BITSET_WORD_BITS] & ((bitset_word_t) 1 << i % BITSET_WORD_BITS))
  90. #define bitset_empty(set) memset (set, '\0', sizeof (bitset_t))
  91. #define bitset_set_all(set) memset (set, '\xff', sizeof (bitset_t))
  92. #define bitset_copy(dest,src) memcpy (dest, src, sizeof (bitset_t))
  93. #define PREV_WORD_CONSTRAINT 0x0001
  94. #define PREV_NOTWORD_CONSTRAINT 0x0002
  95. #define NEXT_WORD_CONSTRAINT 0x0004
  96. #define NEXT_NOTWORD_CONSTRAINT 0x0008
  97. #define PREV_NEWLINE_CONSTRAINT 0x0010
  98. #define NEXT_NEWLINE_CONSTRAINT 0x0020
  99. #define PREV_BEGBUF_CONSTRAINT 0x0040
  100. #define NEXT_ENDBUF_CONSTRAINT 0x0080
  101. #define WORD_DELIM_CONSTRAINT 0x0100
  102. #define NOT_WORD_DELIM_CONSTRAINT 0x0200
  103. typedef enum
  104. {
  105. INSIDE_WORD = PREV_WORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
  106. WORD_FIRST = PREV_NOTWORD_CONSTRAINT | NEXT_WORD_CONSTRAINT,
  107. WORD_LAST = PREV_WORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
  108. INSIDE_NOTWORD = PREV_NOTWORD_CONSTRAINT | NEXT_NOTWORD_CONSTRAINT,
  109. LINE_FIRST = PREV_NEWLINE_CONSTRAINT,
  110. LINE_LAST = NEXT_NEWLINE_CONSTRAINT,
  111. BUF_FIRST = PREV_BEGBUF_CONSTRAINT,
  112. BUF_LAST = NEXT_ENDBUF_CONSTRAINT,
  113. WORD_DELIM = WORD_DELIM_CONSTRAINT,
  114. NOT_WORD_DELIM = NOT_WORD_DELIM_CONSTRAINT
  115. } re_context_type;
  116. typedef struct
  117. {
  118. int alloc;
  119. int nelem;
  120. int *elems;
  121. } re_node_set;
  122. typedef enum
  123. {
  124. NON_TYPE = 0,
  125. /* Node type, These are used by token, node, tree. */
  126. CHARACTER = 1,
  127. END_OF_RE = 2,
  128. SIMPLE_BRACKET = 3,
  129. OP_BACK_REF = 4,
  130. OP_PERIOD = 5,
  131. #ifdef RE_ENABLE_I18N
  132. COMPLEX_BRACKET = 6,
  133. OP_UTF8_PERIOD = 7,
  134. #endif /* RE_ENABLE_I18N */
  135. /* We define EPSILON_BIT as a macro so that OP_OPEN_SUBEXP is used
  136. when the debugger shows values of this enum type. */
  137. #define EPSILON_BIT 8
  138. OP_OPEN_SUBEXP = EPSILON_BIT | 0,
  139. OP_CLOSE_SUBEXP = EPSILON_BIT | 1,
  140. OP_ALT = EPSILON_BIT | 2,
  141. OP_DUP_ASTERISK = EPSILON_BIT | 3,
  142. ANCHOR = EPSILON_BIT | 4,
  143. /* Tree type, these are used only by tree. */
  144. CONCAT = 16,
  145. SUBEXP = 17,
  146. /* Token type, these are used only by token. */
  147. OP_DUP_PLUS = 18,
  148. OP_DUP_QUESTION,
  149. OP_OPEN_BRACKET,
  150. OP_CLOSE_BRACKET,
  151. OP_CHARSET_RANGE,
  152. OP_OPEN_DUP_NUM,
  153. OP_CLOSE_DUP_NUM,
  154. OP_NON_MATCH_LIST,
  155. OP_OPEN_COLL_ELEM,
  156. OP_CLOSE_COLL_ELEM,
  157. OP_OPEN_EQUIV_CLASS,
  158. OP_CLOSE_EQUIV_CLASS,
  159. OP_OPEN_CHAR_CLASS,
  160. OP_CLOSE_CHAR_CLASS,
  161. OP_WORD,
  162. OP_NOTWORD,
  163. OP_SPACE,
  164. OP_NOTSPACE,
  165. BACK_SLASH
  166. } re_token_type_t;
  167. #ifdef RE_ENABLE_I18N
  168. typedef struct
  169. {
  170. /* Multibyte characters. */
  171. wchar_t *mbchars;
  172. /* Collating symbols. */
  173. # if 0
  174. int32_t *coll_syms;
  175. # endif
  176. /* Equivalence classes. */
  177. # if 0
  178. int32_t *equiv_classes;
  179. # endif
  180. /* Range expressions. */
  181. # if 0
  182. uint32_t *range_starts;
  183. uint32_t *range_ends;
  184. # else
  185. wchar_t *range_starts;
  186. wchar_t *range_ends;
  187. # endif
  188. /* Character classes. */
  189. wctype_t *char_classes;
  190. /* If this character set is the non-matching list. */
  191. unsigned int non_match : 1;
  192. /* # of multibyte characters. */
  193. int nmbchars;
  194. /* # of collating symbols. */
  195. int ncoll_syms;
  196. /* # of equivalence classes. */
  197. int nequiv_classes;
  198. /* # of range expressions. */
  199. int nranges;
  200. /* # of character classes. */
  201. int nchar_classes;
  202. } re_charset_t;
  203. #endif /* RE_ENABLE_I18N */
  204. typedef struct
  205. {
  206. union
  207. {
  208. unsigned char c; /* for CHARACTER */
  209. re_bitset_ptr_t sbcset; /* for SIMPLE_BRACKET */
  210. #ifdef RE_ENABLE_I18N
  211. re_charset_t *mbcset; /* for COMPLEX_BRACKET */
  212. #endif /* RE_ENABLE_I18N */
  213. int idx; /* for BACK_REF */
  214. re_context_type ctx_type; /* for ANCHOR */
  215. } opr;
  216. #if __GNUC__ >= 2
  217. re_token_type_t type : 8;
  218. #else
  219. re_token_type_t type;
  220. #endif
  221. unsigned int constraint : 10; /* context constraint */
  222. unsigned int duplicated : 1;
  223. unsigned int opt_subexp : 1;
  224. #ifdef RE_ENABLE_I18N
  225. unsigned int accept_mb : 1;
  226. /* These 2 bits can be moved into the union if needed (e.g. if running out
  227. of bits; move opr.c to opr.c.c and move the flags to opr.c.flags). */
  228. unsigned int mb_partial : 1;
  229. #endif
  230. unsigned int word_char : 1;
  231. } re_token_t;
  232. #define IS_EPSILON_NODE(type) ((type) & EPSILON_BIT)
  233. struct re_string_t
  234. {
  235. /* Indicate the raw buffer which is the original string passed as an
  236. argument of regexec(), re_search(), etc.. */
  237. const unsigned char *raw_mbs;
  238. /* Store the multibyte string. In case of "case insensitive mode" like
  239. REG_ICASE, upper cases of the string are stored, otherwise MBS points
  240. the same address that RAW_MBS points. */
  241. unsigned char *mbs;
  242. #ifdef RE_ENABLE_I18N
  243. /* Store the wide character string which is corresponding to MBS. */
  244. wint_t *wcs;
  245. int *offsets;
  246. mbstate_t cur_state;
  247. #endif
  248. /* Index in RAW_MBS. Each character mbs[i] corresponds to
  249. raw_mbs[raw_mbs_idx + i]. */
  250. int raw_mbs_idx;
  251. /* The length of the valid characters in the buffers. */
  252. int valid_len;
  253. /* The corresponding number of bytes in raw_mbs array. */
  254. int valid_raw_len;
  255. /* The length of the buffers MBS and WCS. */
  256. int bufs_len;
  257. /* The index in MBS, which is updated by re_string_fetch_byte. */
  258. int cur_idx;
  259. /* length of RAW_MBS array. */
  260. int raw_len;
  261. /* This is RAW_LEN - RAW_MBS_IDX + VALID_LEN - VALID_RAW_LEN. */
  262. int len;
  263. /* End of the buffer may be shorter than its length in the cases such
  264. as re_match_2, re_search_2. Then, we use STOP for end of the buffer
  265. instead of LEN. */
  266. int raw_stop;
  267. /* This is RAW_STOP - RAW_MBS_IDX adjusted through OFFSETS. */
  268. int stop;
  269. /* The context of mbs[0]. We store the context independently, since
  270. the context of mbs[0] may be different from raw_mbs[0], which is
  271. the beginning of the input string. */
  272. unsigned int tip_context;
  273. /* The translation passed as a part of an argument of re_compile_pattern. */
  274. RE_TRANSLATE_TYPE trans;
  275. /* Copy of re_dfa_t's word_char. */
  276. re_const_bitset_ptr_t word_char;
  277. /* 1 if REG_ICASE. */
  278. unsigned char icase;
  279. unsigned char is_utf8;
  280. unsigned char map_notascii;
  281. unsigned char mbs_allocated;
  282. unsigned char offsets_needed;
  283. unsigned char newline_anchor;
  284. unsigned char word_ops_used;
  285. int mb_cur_max;
  286. };
  287. typedef struct re_string_t re_string_t;
  288. struct re_dfa_t;
  289. typedef struct re_dfa_t re_dfa_t;
  290. static reg_errcode_t re_string_realloc_buffers (re_string_t *pstr,
  291. int new_buf_len)
  292. internal_function;
  293. #ifdef RE_ENABLE_I18N
  294. static void build_wcs_buffer (re_string_t *pstr) internal_function;
  295. static reg_errcode_t build_wcs_upper_buffer (re_string_t *pstr) internal_function;
  296. #endif /* RE_ENABLE_I18N */
  297. static void build_upper_buffer (re_string_t *pstr) internal_function;
  298. static void re_string_translate_buffer (re_string_t *pstr) internal_function;
  299. static unsigned int re_string_context_at (const re_string_t *input, int idx,
  300. int eflags)
  301. internal_function __attribute ((pure));
  302. #define re_string_peek_byte(pstr, offset) \
  303. ((pstr)->mbs[(pstr)->cur_idx + offset])
  304. #define re_string_fetch_byte(pstr) \
  305. ((pstr)->mbs[(pstr)->cur_idx++])
  306. #define re_string_first_byte(pstr, idx) \
  307. ((idx) == (pstr)->valid_len || (pstr)->wcs[idx] != WEOF)
  308. #define re_string_is_single_byte_char(pstr, idx) \
  309. ((pstr)->wcs[idx] != WEOF && ((pstr)->valid_len == (idx) + 1 \
  310. || (pstr)->wcs[(idx) + 1] != WEOF))
  311. #define re_string_eoi(pstr) ((pstr)->stop <= (pstr)->cur_idx)
  312. #define re_string_cur_idx(pstr) ((pstr)->cur_idx)
  313. #define re_string_get_buffer(pstr) ((pstr)->mbs)
  314. #define re_string_length(pstr) ((pstr)->len)
  315. #define re_string_byte_at(pstr,idx) ((pstr)->mbs[idx])
  316. #define re_string_skip_bytes(pstr,idx) ((pstr)->cur_idx += (idx))
  317. #define re_string_set_index(pstr,idx) ((pstr)->cur_idx = (idx))
  318. #include <alloca.h>
  319. #if 1
  320. # ifdef HAVE_ALLOCA
  321. /* The OS usually guarantees only one guard page at the bottom of the stack,
  322. and a page size can be as small as 4096 bytes. So we cannot safely
  323. allocate anything larger than 4096 bytes. Also care for the possibility
  324. of a few compiler-allocated temporary stack slots. */
  325. # define __libc_use_alloca(n) ((n) < 4032)
  326. # else
  327. /* alloca is implemented with malloc, so just use malloc. */
  328. # define __libc_use_alloca(n) 0
  329. # endif
  330. #endif
  331. #define re_malloc(t,n) ((t *) malloc ((n) * sizeof (t)))
  332. #define re_realloc(p,t,n) ((t *) realloc (p, (n) * sizeof (t)))
  333. #define re_free(p) free (p)
  334. struct bin_tree_t
  335. {
  336. struct bin_tree_t *parent;
  337. struct bin_tree_t *left;
  338. struct bin_tree_t *right;
  339. struct bin_tree_t *first;
  340. struct bin_tree_t *next;
  341. re_token_t token;
  342. /* `node_idx' is the index in dfa->nodes, if `type' == 0.
  343. Otherwise `type' indicate the type of this node. */
  344. int node_idx;
  345. };
  346. typedef struct bin_tree_t bin_tree_t;
  347. #define BIN_TREE_STORAGE_SIZE \
  348. ((1024 - sizeof (void *)) / sizeof (bin_tree_t))
  349. struct bin_tree_storage_t
  350. {
  351. struct bin_tree_storage_t *next;
  352. bin_tree_t data[BIN_TREE_STORAGE_SIZE];
  353. };
  354. typedef struct bin_tree_storage_t bin_tree_storage_t;
  355. #define CONTEXT_WORD 1
  356. #define CONTEXT_NEWLINE (CONTEXT_WORD << 1)
  357. #define CONTEXT_BEGBUF (CONTEXT_NEWLINE << 1)
  358. #define CONTEXT_ENDBUF (CONTEXT_BEGBUF << 1)
  359. #define IS_WORD_CONTEXT(c) ((c) & CONTEXT_WORD)
  360. #define IS_NEWLINE_CONTEXT(c) ((c) & CONTEXT_NEWLINE)
  361. #define IS_BEGBUF_CONTEXT(c) ((c) & CONTEXT_BEGBUF)
  362. #define IS_ENDBUF_CONTEXT(c) ((c) & CONTEXT_ENDBUF)
  363. #define IS_ORDINARY_CONTEXT(c) ((c) == 0)
  364. #define IS_WORD_CHAR(ch) (isalnum (ch) || (ch) == '_')
  365. #define IS_NEWLINE(ch) ((ch) == NEWLINE_CHAR)
  366. #define IS_WIDE_WORD_CHAR(ch) (iswalnum (ch) || (ch) == L'_')
  367. #define IS_WIDE_NEWLINE(ch) ((ch) == WIDE_NEWLINE_CHAR)
  368. #define NOT_SATISFY_PREV_CONSTRAINT(constraint,context) \
  369. ((((constraint) & PREV_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
  370. || ((constraint & PREV_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
  371. || ((constraint & PREV_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context))\
  372. || ((constraint & PREV_BEGBUF_CONSTRAINT) && !IS_BEGBUF_CONTEXT (context)))
  373. #define NOT_SATISFY_NEXT_CONSTRAINT(constraint,context) \
  374. ((((constraint) & NEXT_WORD_CONSTRAINT) && !IS_WORD_CONTEXT (context)) \
  375. || (((constraint) & NEXT_NOTWORD_CONSTRAINT) && IS_WORD_CONTEXT (context)) \
  376. || (((constraint) & NEXT_NEWLINE_CONSTRAINT) && !IS_NEWLINE_CONTEXT (context)) \
  377. || (((constraint) & NEXT_ENDBUF_CONSTRAINT) && !IS_ENDBUF_CONTEXT (context)))
  378. struct re_dfastate_t
  379. {
  380. unsigned int hash;
  381. re_node_set nodes;
  382. re_node_set non_eps_nodes;
  383. re_node_set inveclosure;
  384. re_node_set *entrance_nodes;
  385. struct re_dfastate_t **trtable, **word_trtable;
  386. unsigned int context : 4;
  387. unsigned int halt : 1;
  388. /* If this state can accept `multi byte'.
  389. Note that we refer to multibyte characters, and multi character
  390. collating elements as `multi byte'. */
  391. unsigned int accept_mb : 1;
  392. /* If this state has backreference node(s). */
  393. unsigned int has_backref : 1;
  394. unsigned int has_constraint : 1;
  395. };
  396. typedef struct re_dfastate_t re_dfastate_t;
  397. struct re_state_table_entry
  398. {
  399. int num;
  400. int alloc;
  401. re_dfastate_t **array;
  402. };
  403. /* Array type used in re_sub_match_last_t and re_sub_match_top_t. */
  404. typedef struct
  405. {
  406. int next_idx;
  407. int alloc;
  408. re_dfastate_t **array;
  409. } state_array_t;
  410. /* Store information about the node NODE whose type is OP_CLOSE_SUBEXP. */
  411. typedef struct
  412. {
  413. int node;
  414. int str_idx; /* The position NODE match at. */
  415. state_array_t path;
  416. } re_sub_match_last_t;
  417. /* Store information about the node NODE whose type is OP_OPEN_SUBEXP.
  418. And information about the node, whose type is OP_CLOSE_SUBEXP,
  419. corresponding to NODE is stored in LASTS. */
  420. typedef struct
  421. {
  422. int str_idx;
  423. int node;
  424. state_array_t *path;
  425. int alasts; /* Allocation size of LASTS. */
  426. int nlasts; /* The number of LASTS. */
  427. re_sub_match_last_t **lasts;
  428. } re_sub_match_top_t;
  429. struct re_backref_cache_entry
  430. {
  431. int node;
  432. int str_idx;
  433. int subexp_from;
  434. int subexp_to;
  435. char more;
  436. char unused;
  437. unsigned short int eps_reachable_subexps_map;
  438. };
  439. typedef struct
  440. {
  441. /* The string object corresponding to the input string. */
  442. re_string_t input;
  443. const re_dfa_t *dfa;
  444. /* EFLAGS of the argument of regexec. */
  445. int eflags;
  446. /* Where the matching ends. */
  447. int match_last;
  448. int last_node;
  449. /* The state log used by the matcher. */
  450. re_dfastate_t **state_log;
  451. int state_log_top;
  452. /* Back reference cache. */
  453. int nbkref_ents;
  454. int abkref_ents;
  455. struct re_backref_cache_entry *bkref_ents;
  456. int max_mb_elem_len;
  457. int nsub_tops;
  458. int asub_tops;
  459. re_sub_match_top_t **sub_tops;
  460. } re_match_context_t;
  461. typedef struct
  462. {
  463. re_dfastate_t **sifted_states;
  464. re_dfastate_t **limited_states;
  465. int last_node;
  466. int last_str_idx;
  467. re_node_set limits;
  468. } re_sift_context_t;
  469. struct re_fail_stack_ent_t
  470. {
  471. int idx;
  472. int node;
  473. regmatch_t *regs;
  474. re_node_set eps_via_nodes;
  475. };
  476. struct re_fail_stack_t
  477. {
  478. int num;
  479. int alloc;
  480. struct re_fail_stack_ent_t *stack;
  481. };
  482. struct re_dfa_t
  483. {
  484. re_token_t *nodes;
  485. size_t nodes_alloc;
  486. size_t nodes_len;
  487. int *nexts;
  488. int *org_indices;
  489. re_node_set *edests;
  490. re_node_set *eclosures;
  491. re_node_set *inveclosures;
  492. struct re_state_table_entry *state_table;
  493. re_dfastate_t *init_state;
  494. re_dfastate_t *init_state_word;
  495. re_dfastate_t *init_state_nl;
  496. re_dfastate_t *init_state_begbuf;
  497. bin_tree_t *str_tree;
  498. bin_tree_storage_t *str_tree_storage;
  499. re_bitset_ptr_t sb_char;
  500. int str_tree_storage_idx;
  501. /* number of subexpressions `re_nsub' is in regex_t. */
  502. unsigned int state_hash_mask;
  503. int init_node;
  504. int nbackref; /* The number of backreference in this dfa. */
  505. /* Bitmap expressing which backreference is used. */
  506. bitset_word_t used_bkref_map;
  507. bitset_word_t completed_bkref_map;
  508. unsigned int has_plural_match : 1;
  509. /* If this dfa has "multibyte node", which is a backreference or
  510. a node which can accept multibyte character or multi character
  511. collating element. */
  512. unsigned int has_mb_node : 1;
  513. unsigned int is_utf8 : 1;
  514. unsigned int map_notascii : 1;
  515. unsigned int word_ops_used : 1;
  516. int mb_cur_max;
  517. bitset_t word_char;
  518. reg_syntax_t syntax;
  519. int *subexp_map;
  520. #ifdef DEBUG
  521. char* re_str;
  522. #endif
  523. __libc_lock_define (, lock)
  524. };
  525. #define re_node_set_init_empty(set) memset (set, '\0', sizeof (re_node_set))
  526. #define re_node_set_remove(set,id) \
  527. (re_node_set_remove_at (set, re_node_set_contains (set, id) - 1))
  528. #define re_node_set_empty(p) ((p)->nelem = 0)
  529. #define re_node_set_free(set) re_free ((set)->elems)
  530. typedef enum
  531. {
  532. SB_CHAR,
  533. MB_CHAR,
  534. EQUIV_CLASS,
  535. COLL_SYM,
  536. CHAR_CLASS
  537. } bracket_elem_type;
  538. typedef struct
  539. {
  540. bracket_elem_type type;
  541. union
  542. {
  543. unsigned char ch;
  544. unsigned char *name;
  545. #ifdef __UCLIBC_HAS_WCHAR__
  546. wchar_t wch;
  547. #endif
  548. } opr;
  549. } bracket_elem_t;
  550. /* Inline functions for bitset operation. */
  551. static __inline__ void
  552. bitset_not (bitset_t set)
  553. {
  554. int bitset_i;
  555. for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
  556. set[bitset_i] = ~set[bitset_i];
  557. }
  558. static __inline__ void
  559. bitset_merge (bitset_t dest, const bitset_t src)
  560. {
  561. int bitset_i;
  562. for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
  563. dest[bitset_i] |= src[bitset_i];
  564. }
  565. static __inline__ void
  566. bitset_mask (bitset_t dest, const bitset_t src)
  567. {
  568. int bitset_i;
  569. for (bitset_i = 0; bitset_i < BITSET_WORDS; ++bitset_i)
  570. dest[bitset_i] &= src[bitset_i];
  571. }
  572. #ifdef RE_ENABLE_I18N
  573. /* Inline functions for re_string. */
  574. static __inline__ int
  575. internal_function __attribute ((pure))
  576. re_string_char_size_at (const re_string_t *pstr, int idx)
  577. {
  578. int byte_idx;
  579. if (pstr->mb_cur_max == 1)
  580. return 1;
  581. for (byte_idx = 1; idx + byte_idx < pstr->valid_len; ++byte_idx)
  582. if (pstr->wcs[idx + byte_idx] != WEOF)
  583. break;
  584. return byte_idx;
  585. }
  586. static __inline__ wint_t
  587. internal_function __attribute ((pure))
  588. re_string_wchar_at (const re_string_t *pstr, int idx)
  589. {
  590. if (pstr->mb_cur_max == 1)
  591. return (wint_t) pstr->mbs[idx];
  592. return (wint_t) pstr->wcs[idx];
  593. }
  594. static int
  595. internal_function __attribute ((pure))
  596. re_string_elem_size_at (const re_string_t *pstr, int idx)
  597. {
  598. # if 0
  599. const unsigned char *p, *extra;
  600. const int32_t *table, *indirect;
  601. int32_t tmp;
  602. # include <locale/weight.h>
  603. uint_fast32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
  604. if (nrules != 0)
  605. {
  606. table = (const int32_t *) _NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
  607. extra = (const unsigned char *)
  608. _NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
  609. indirect = (const int32_t *) _NL_CURRENT (LC_COLLATE,
  610. _NL_COLLATE_INDIRECTMB);
  611. p = pstr->mbs + idx;
  612. tmp = findidx (&p);
  613. return p - pstr->mbs - idx;
  614. }
  615. # endif
  616. return 1;
  617. }
  618. #endif /* RE_ENABLE_I18N */
  619. #endif /* _REGEX_INTERNAL_H */