| File: | kwset.c |
| Location: | line 462, column 19 |
| Description: | Access to field 'next' results in a dereference of a null pointer (loaded from field 'trie') |
| 1 | /* | |||
| 2 | * This file has been copied from commit e7ac713d^ in the GNU grep git | |||
| 3 | * repository. A few small changes have been made to adapt the code to | |||
| 4 | * Git. | |||
| 5 | */ | |||
| 6 | ||||
| 7 | /* kwset.c - search for any of a set of keywords. | |||
| 8 | Copyright 1989, 1998, 2000, 2005 Free Software Foundation, Inc. | |||
| 9 | ||||
| 10 | This program is free software; you can redistribute it and/or modify | |||
| 11 | it under the terms of the GNU General Public License as published by | |||
| 12 | the Free Software Foundation; either version 2, or (at your option) | |||
| 13 | any later version. | |||
| 14 | ||||
| 15 | This program is distributed in the hope that it will be useful, | |||
| 16 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |||
| 17 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |||
| 18 | GNU General Public License for more details. | |||
| 19 | ||||
| 20 | You should have received a copy of the GNU General Public License | |||
| 21 | along with this program; if not, write to the Free Software | |||
| 22 | Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, MA | |||
| 23 | 02110-1301, USA. */ | |||
| 24 | ||||
| 25 | /* Written August 1989 by Mike Haertel. | |||
| 26 | The author may be reached (Email) at the address mike@ai.mit.edu, | |||
| 27 | or (US mail) as Mike Haertel c/o Free Software Foundation. */ | |||
| 28 | ||||
| 29 | /* The algorithm implemented by these routines bears a startling resemblance | |||
| 30 | to one discovered by Beate Commentz-Walter, although it is not identical. | |||
| 31 | See "A String Matching Algorithm Fast on the Average," Technical Report, | |||
| 32 | IBM-Germany, Scientific Center Heidelberg, Tiergartenstrasse 15, D-6900 | |||
| 33 | Heidelberg, Germany. See also Aho, A.V., and M. Corasick, "Efficient | |||
| 34 | String Matching: An Aid to Bibliographic Search," CACM June 1975, | |||
| 35 | Vol. 18, No. 6, which describes the failure function used below. */ | |||
| 36 | ||||
| 37 | #include "cache.h" | |||
| 38 | ||||
| 39 | #include "kwset.h" | |||
| 40 | #include "compat/obstack.h" | |||
| 41 | ||||
| 42 | #define NCHAR((127*2 +1) + 1) (UCHAR_MAX(127*2 +1) + 1) | |||
| 43 | #define obstack_chunk_allocxmalloc xmalloc | |||
| 44 | #define obstack_chunk_freefree free | |||
| 45 | ||||
| 46 | #define U(c)((unsigned char) (c)) ((unsigned char) (c)) | |||
| 47 | ||||
| 48 | /* Balanced tree of edges and labels leaving a given trie node. */ | |||
| 49 | struct tree | |||
| 50 | { | |||
| 51 | struct tree *llink; /* Left link; MUST be first field. */ | |||
| 52 | struct tree *rlink; /* Right link (to larger labels). */ | |||
| 53 | struct trie *trie; /* Trie node pointed to by this edge. */ | |||
| 54 | unsigned char label; /* Label on this edge. */ | |||
| 55 | char balance; /* Difference in depths of subtrees. */ | |||
| 56 | }; | |||
| 57 | ||||
| 58 | /* Node of a trie representing a set of reversed keywords. */ | |||
| 59 | struct trie | |||
| 60 | { | |||
| 61 | unsigned int accepting; /* Word index of accepted word, or zero. */ | |||
| 62 | struct tree *links; /* Tree of edges leaving this node. */ | |||
| 63 | struct trie *parent; /* Parent of this node. */ | |||
| 64 | struct trie *next; /* List of all trie nodes in level order. */ | |||
| 65 | struct trie *fail; /* Aho-Corasick failure function. */ | |||
| 66 | int depth; /* Depth of this node from the root. */ | |||
| 67 | int shift; /* Shift function for search failures. */ | |||
| 68 | int maxshift; /* Max shift of self and descendants. */ | |||
| 69 | }; | |||
| 70 | ||||
| 71 | /* Structure returned opaquely to the caller, containing everything. */ | |||
| 72 | struct kwset | |||
| 73 | { | |||
| 74 | struct obstack obstack; /* Obstack for node allocation. */ | |||
| 75 | int words; /* Number of words in the trie. */ | |||
| 76 | struct trie *trie; /* The trie itself. */ | |||
| 77 | int mind; /* Minimum depth of an accepting node. */ | |||
| 78 | int maxd; /* Maximum depth of any node. */ | |||
| 79 | unsigned char delta[NCHAR((127*2 +1) + 1)]; /* Delta table for rapid search. */ | |||
| 80 | struct trie *next[NCHAR((127*2 +1) + 1)]; /* Table of children of the root. */ | |||
| 81 | char *target; /* Target string if there's only one. */ | |||
| 82 | int mind2; /* Used in Boyer-Moore search for one string. */ | |||
| 83 | unsigned char const *trans; /* Character translation table. */ | |||
| 84 | }; | |||
| 85 | ||||
| 86 | /* Allocate and initialize a keyword set object, returning an opaque | |||
| 87 | pointer to it. Return NULL if memory is not available. */ | |||
| 88 | kwset_t | |||
| 89 | kwsalloc (unsigned char const *trans) | |||
| 90 | { | |||
| 91 | struct kwset *kwset; | |||
| 92 | ||||
| 93 | kwset = (struct kwset *) xmalloc(sizeof (struct kwset)); | |||
| 94 | ||||
| 95 | obstack_init(&kwset->obstack)_obstack_begin ((&kwset->obstack), 0, 0, (void *(*) (long )) xmalloc, (void (*) (void *)) free); | |||
| 96 | kwset->words = 0; | |||
| 97 | kwset->trie | |||
| 98 | = (struct trie *) obstack_alloc(&kwset->obstack, sizeof (struct trie))__extension__ ({ struct obstack *__h = (&kwset->obstack ); __extension__ ({ struct obstack *__o = (__h); int __len = ( (sizeof (struct trie))); if (__o->chunk_limit - __o->next_free < __len) _obstack_newchunk (__o, __len); ((__o)->next_free += (__len)); (void) 0; }); __extension__ ({ struct obstack * __o1 = (__h); void *__value = (void *) __o1->object_base; if (__o1->next_free == __value) __o1->maybe_empty_object = 1; __o1->next_free = ((sizeof (long int) < sizeof (void *) ? (__o1->object_base) : (char *) 0) + (((__o1->next_free ) - (sizeof (long int) < sizeof (void *) ? (__o1->object_base ) : (char *) 0) + (__o1->alignment_mask)) & ~(__o1-> alignment_mask))); if (__o1->next_free - (char *)__o1-> chunk > __o1->chunk_limit - (char *)__o1->chunk) __o1 ->next_free = __o1->chunk_limit; __o1->object_base = __o1->next_free; __value; }); }); | |||
| 99 | if (!kwset->trie) | |||
| 100 | { | |||
| 101 | kwsfree((kwset_t) kwset); | |||
| 102 | return NULL((void*)0); | |||
| 103 | } | |||
| 104 | kwset->trie->accepting = 0; | |||
| 105 | kwset->trie->links = NULL((void*)0); | |||
| 106 | kwset->trie->parent = NULL((void*)0); | |||
| 107 | kwset->trie->next = NULL((void*)0); | |||
| 108 | kwset->trie->fail = NULL((void*)0); | |||
| 109 | kwset->trie->depth = 0; | |||
| 110 | kwset->trie->shift = 0; | |||
| 111 | kwset->mind = INT_MAX2147483647; | |||
| 112 | kwset->maxd = -1; | |||
| 113 | kwset->target = NULL((void*)0); | |||
| 114 | kwset->trans = trans; | |||
| 115 | ||||
| 116 | return (kwset_t) kwset; | |||
| 117 | } | |||
| 118 | ||||
| 119 | /* This upper bound is valid for CHAR_BIT >= 4 and | |||
| 120 | exact for CHAR_BIT in { 4..11, 13, 15, 17, 19 }. */ | |||
| 121 | #define DEPTH_SIZE(8 + 8/2) (CHAR_BIT8 + CHAR_BIT8/2) | |||
| 122 | ||||
| 123 | /* Add the given string to the contents of the keyword set. Return NULL | |||
| 124 | for success, an error message otherwise. */ | |||
| 125 | const char * | |||
| 126 | kwsincr (kwset_t kws, char const *text, size_t len) | |||
| 127 | { | |||
| 128 | struct kwset *kwset; | |||
| 129 | register struct trie *trie; | |||
| 130 | register unsigned char label; | |||
| 131 | register struct tree *link; | |||
| 132 | register int depth; | |||
| 133 | struct tree *links[DEPTH_SIZE(8 + 8/2)]; | |||
| 134 | enum { L, R } dirs[DEPTH_SIZE(8 + 8/2)]; | |||
| 135 | struct tree *t, *r, *l, *rl, *lr; | |||
| 136 | ||||
| 137 | kwset = (struct kwset *) kws; | |||
| 138 | trie = kwset->trie; | |||
| 139 | text += len; | |||
| 140 | ||||
| 141 | /* Descend the trie (built of reversed keywords) character-by-character, | |||
| 142 | installing new nodes when necessary. */ | |||
| 143 | while (len--) | |||
| 144 | { | |||
| 145 | label = kwset->trans ? kwset->trans[U(*--text)((unsigned char) (*--text))] : *--text; | |||
| 146 | ||||
| 147 | /* Descend the tree of outgoing links for this trie node, | |||
| 148 | looking for the current character and keeping track | |||
| 149 | of the path followed. */ | |||
| 150 | link = trie->links; | |||
| 151 | links[0] = (struct tree *) &trie->links; | |||
| 152 | dirs[0] = L; | |||
| 153 | depth = 1; | |||
| 154 | ||||
| 155 | while (link && label != link->label) | |||
| 156 | { | |||
| 157 | links[depth] = link; | |||
| 158 | if (label < link->label) | |||
| 159 | dirs[depth++] = L, link = link->llink; | |||
| 160 | else | |||
| 161 | dirs[depth++] = R, link = link->rlink; | |||
| 162 | } | |||
| 163 | ||||
| 164 | /* The current character doesn't have an outgoing link at | |||
| 165 | this trie node, so build a new trie node and install | |||
| 166 | a link in the current trie node's tree. */ | |||
| 167 | if (!link) | |||
| 168 | { | |||
| 169 | link = (struct tree *) obstack_alloc(&kwset->obstack,__extension__ ({ struct obstack *__h = (&kwset->obstack ); __extension__ ({ struct obstack *__o = (__h); int __len = ( (sizeof (struct tree))); if (__o->chunk_limit - __o->next_free < __len) _obstack_newchunk (__o, __len); ((__o)->next_free += (__len)); (void) 0; }); __extension__ ({ struct obstack * __o1 = (__h); void *__value = (void *) __o1->object_base; if (__o1->next_free == __value) __o1->maybe_empty_object = 1; __o1->next_free = ((sizeof (long int) < sizeof (void *) ? (__o1->object_base) : (char *) 0) + (((__o1->next_free ) - (sizeof (long int) < sizeof (void *) ? (__o1->object_base ) : (char *) 0) + (__o1->alignment_mask)) & ~(__o1-> alignment_mask))); if (__o1->next_free - (char *)__o1-> chunk > __o1->chunk_limit - (char *)__o1->chunk) __o1 ->next_free = __o1->chunk_limit; __o1->object_base = __o1->next_free; __value; }); }) | |||
| 170 | sizeof (struct tree))__extension__ ({ struct obstack *__h = (&kwset->obstack ); __extension__ ({ struct obstack *__o = (__h); int __len = ( (sizeof (struct tree))); if (__o->chunk_limit - __o->next_free < __len) _obstack_newchunk (__o, __len); ((__o)->next_free += (__len)); (void) 0; }); __extension__ ({ struct obstack * __o1 = (__h); void *__value = (void *) __o1->object_base; if (__o1->next_free == __value) __o1->maybe_empty_object = 1; __o1->next_free = ((sizeof (long int) < sizeof (void *) ? (__o1->object_base) : (char *) 0) + (((__o1->next_free ) - (sizeof (long int) < sizeof (void *) ? (__o1->object_base ) : (char *) 0) + (__o1->alignment_mask)) & ~(__o1-> alignment_mask))); if (__o1->next_free - (char *)__o1-> chunk > __o1->chunk_limit - (char *)__o1->chunk) __o1 ->next_free = __o1->chunk_limit; __o1->object_base = __o1->next_free; __value; }); }); | |||
| 171 | if (!link) | |||
| 172 | return "memory exhausted"; | |||
| 173 | link->llink = NULL((void*)0); | |||
| 174 | link->rlink = NULL((void*)0); | |||
| 175 | link->trie = (struct trie *) obstack_alloc(&kwset->obstack,__extension__ ({ struct obstack *__h = (&kwset->obstack ); __extension__ ({ struct obstack *__o = (__h); int __len = ( (sizeof (struct trie))); if (__o->chunk_limit - __o->next_free < __len) _obstack_newchunk (__o, __len); ((__o)->next_free += (__len)); (void) 0; }); __extension__ ({ struct obstack * __o1 = (__h); void *__value = (void *) __o1->object_base; if (__o1->next_free == __value) __o1->maybe_empty_object = 1; __o1->next_free = ((sizeof (long int) < sizeof (void *) ? (__o1->object_base) : (char *) 0) + (((__o1->next_free ) - (sizeof (long int) < sizeof (void *) ? (__o1->object_base ) : (char *) 0) + (__o1->alignment_mask)) & ~(__o1-> alignment_mask))); if (__o1->next_free - (char *)__o1-> chunk > __o1->chunk_limit - (char *)__o1->chunk) __o1 ->next_free = __o1->chunk_limit; __o1->object_base = __o1->next_free; __value; }); }) | |||
| 176 | sizeof (struct trie))__extension__ ({ struct obstack *__h = (&kwset->obstack ); __extension__ ({ struct obstack *__o = (__h); int __len = ( (sizeof (struct trie))); if (__o->chunk_limit - __o->next_free < __len) _obstack_newchunk (__o, __len); ((__o)->next_free += (__len)); (void) 0; }); __extension__ ({ struct obstack * __o1 = (__h); void *__value = (void *) __o1->object_base; if (__o1->next_free == __value) __o1->maybe_empty_object = 1; __o1->next_free = ((sizeof (long int) < sizeof (void *) ? (__o1->object_base) : (char *) 0) + (((__o1->next_free ) - (sizeof (long int) < sizeof (void *) ? (__o1->object_base ) : (char *) 0) + (__o1->alignment_mask)) & ~(__o1-> alignment_mask))); if (__o1->next_free - (char *)__o1-> chunk > __o1->chunk_limit - (char *)__o1->chunk) __o1 ->next_free = __o1->chunk_limit; __o1->object_base = __o1->next_free; __value; }); }); | |||
| 177 | if (!link->trie) | |||
| 178 | { | |||
| 179 | obstack_free(&kwset->obstack, link)__extension__ ({ struct obstack *__o = (&kwset->obstack ); void *__obj = (link); if (__obj > (void *)__o->chunk && __obj < (void *)__o->chunk_limit) __o->next_free = __o->object_base = (char *)__obj; else (obstack_free) ( __o, __obj); }); | |||
| 180 | return "memory exhausted"; | |||
| 181 | } | |||
| 182 | link->trie->accepting = 0; | |||
| 183 | link->trie->links = NULL((void*)0); | |||
| 184 | link->trie->parent = trie; | |||
| 185 | link->trie->next = NULL((void*)0); | |||
| 186 | link->trie->fail = NULL((void*)0); | |||
| 187 | link->trie->depth = trie->depth + 1; | |||
| 188 | link->trie->shift = 0; | |||
| 189 | link->label = label; | |||
| 190 | link->balance = 0; | |||
| 191 | ||||
| 192 | /* Install the new tree node in its parent. */ | |||
| 193 | if (dirs[--depth] == L) | |||
| 194 | links[depth]->llink = link; | |||
| 195 | else | |||
| 196 | links[depth]->rlink = link; | |||
| 197 | ||||
| 198 | /* Back up the tree fixing the balance flags. */ | |||
| 199 | while (depth && !links[depth]->balance) | |||
| 200 | { | |||
| 201 | if (dirs[depth] == L) | |||
| 202 | --links[depth]->balance; | |||
| 203 | else | |||
| 204 | ++links[depth]->balance; | |||
| 205 | --depth; | |||
| 206 | } | |||
| 207 | ||||
| 208 | /* Rebalance the tree by pointer rotations if necessary. */ | |||
| 209 | if (depth && ((dirs[depth] == L && --links[depth]->balance) | |||
| 210 | || (dirs[depth] == R && ++links[depth]->balance))) | |||
| 211 | { | |||
| 212 | switch (links[depth]->balance) | |||
| 213 | { | |||
| 214 | case (char) -2: | |||
| 215 | switch (dirs[depth + 1]) | |||
| 216 | { | |||
| 217 | case L: | |||
| 218 | r = links[depth], t = r->llink, rl = t->rlink; | |||
| 219 | t->rlink = r, r->llink = rl; | |||
| 220 | t->balance = r->balance = 0; | |||
| 221 | break; | |||
| 222 | case R: | |||
| 223 | r = links[depth], l = r->llink, t = l->rlink; | |||
| 224 | rl = t->rlink, lr = t->llink; | |||
| 225 | t->llink = l, l->rlink = lr, t->rlink = r, r->llink = rl; | |||
| 226 | l->balance = t->balance != 1 ? 0 : -1; | |||
| 227 | r->balance = t->balance != (char) -1 ? 0 : 1; | |||
| 228 | t->balance = 0; | |||
| 229 | break; | |||
| 230 | default: | |||
| 231 | abort (); | |||
| 232 | } | |||
| 233 | break; | |||
| 234 | case 2: | |||
| 235 | switch (dirs[depth + 1]) | |||
| 236 | { | |||
| 237 | case R: | |||
| 238 | l = links[depth], t = l->rlink, lr = t->llink; | |||
| 239 | t->llink = l, l->rlink = lr; | |||
| 240 | t->balance = l->balance = 0; | |||
| 241 | break; | |||
| 242 | case L: | |||
| 243 | l = links[depth], r = l->rlink, t = r->llink; | |||
| 244 | lr = t->llink, rl = t->rlink; | |||
| 245 | t->llink = l, l->rlink = lr, t->rlink = r, r->llink = rl; | |||
| 246 | l->balance = t->balance != 1 ? 0 : -1; | |||
| 247 | r->balance = t->balance != (char) -1 ? 0 : 1; | |||
| 248 | t->balance = 0; | |||
| 249 | break; | |||
| 250 | default: | |||
| 251 | abort (); | |||
| 252 | } | |||
| 253 | break; | |||
| 254 | default: | |||
| 255 | abort (); | |||
| 256 | } | |||
| 257 | ||||
| 258 | if (dirs[depth - 1] == L) | |||
| 259 | links[depth - 1]->llink = t; | |||
| 260 | else | |||
| 261 | links[depth - 1]->rlink = t; | |||
| 262 | } | |||
| 263 | } | |||
| 264 | ||||
| 265 | trie = link->trie; | |||
| 266 | } | |||
| 267 | ||||
| 268 | /* Mark the node we finally reached as accepting, encoding the | |||
| 269 | index number of this word in the keyword set so far. */ | |||
| 270 | if (!trie->accepting) | |||
| 271 | trie->accepting = 1 + 2 * kwset->words; | |||
| 272 | ++kwset->words; | |||
| 273 | ||||
| 274 | /* Keep track of the longest and shortest string of the keyword set. */ | |||
| 275 | if (trie->depth < kwset->mind) | |||
| 276 | kwset->mind = trie->depth; | |||
| 277 | if (trie->depth > kwset->maxd) | |||
| 278 | kwset->maxd = trie->depth; | |||
| 279 | ||||
| 280 | return NULL((void*)0); | |||
| 281 | } | |||
| 282 | ||||
| 283 | /* Enqueue the trie nodes referenced from the given tree in the | |||
| 284 | given queue. */ | |||
| 285 | static void | |||
| 286 | enqueue (struct tree *tree, struct trie **last) | |||
| 287 | { | |||
| 288 | if (!tree) | |||
| 289 | return; | |||
| 290 | enqueue(tree->llink, last); | |||
| 291 | enqueue(tree->rlink, last); | |||
| 292 | (*last) = (*last)->next = tree->trie; | |||
| 293 | } | |||
| 294 | ||||
| 295 | /* Compute the Aho-Corasick failure function for the trie nodes referenced | |||
| 296 | from the given tree, given the failure function for their parent as | |||
| 297 | well as a last resort failure node. */ | |||
| 298 | static void | |||
| 299 | treefails (register struct tree const *tree, struct trie const *fail, | |||
| 300 | struct trie *recourse) | |||
| 301 | { | |||
| 302 | register struct tree *link; | |||
| 303 | ||||
| 304 | if (!tree) | |||
| 305 | return; | |||
| 306 | ||||
| 307 | treefails(tree->llink, fail, recourse); | |||
| 308 | treefails(tree->rlink, fail, recourse); | |||
| 309 | ||||
| 310 | /* Find, in the chain of fails going back to the root, the first | |||
| 311 | node that has a descendant on the current label. */ | |||
| 312 | while (fail) | |||
| 313 | { | |||
| 314 | link = fail->links; | |||
| 315 | while (link && tree->label != link->label) | |||
| 316 | if (tree->label < link->label) | |||
| 317 | link = link->llink; | |||
| 318 | else | |||
| 319 | link = link->rlink; | |||
| 320 | if (link) | |||
| 321 | { | |||
| 322 | tree->trie->fail = link->trie; | |||
| 323 | return; | |||
| 324 | } | |||
| 325 | fail = fail->fail; | |||
| 326 | } | |||
| 327 | ||||
| 328 | tree->trie->fail = recourse; | |||
| 329 | } | |||
| 330 | ||||
| 331 | /* Set delta entries for the links of the given tree such that | |||
| 332 | the preexisting delta value is larger than the current depth. */ | |||
| 333 | static void | |||
| 334 | treedelta (register struct tree const *tree, | |||
| 335 | register unsigned int depth, | |||
| 336 | unsigned char delta[]) | |||
| 337 | { | |||
| 338 | if (!tree) | |||
| 339 | return; | |||
| 340 | treedelta(tree->llink, depth, delta); | |||
| 341 | treedelta(tree->rlink, depth, delta); | |||
| 342 | if (depth < delta[tree->label]) | |||
| 343 | delta[tree->label] = depth; | |||
| 344 | } | |||
| 345 | ||||
| 346 | /* Return true if A has every label in B. */ | |||
| 347 | static int | |||
| 348 | hasevery (register struct tree const *a, register struct tree const *b) | |||
| 349 | { | |||
| 350 | if (!b) | |||
| 351 | return 1; | |||
| 352 | if (!hasevery(a, b->llink)) | |||
| 353 | return 0; | |||
| 354 | if (!hasevery(a, b->rlink)) | |||
| 355 | return 0; | |||
| 356 | while (a && b->label != a->label) | |||
| 357 | if (b->label < a->label) | |||
| 358 | a = a->llink; | |||
| 359 | else | |||
| 360 | a = a->rlink; | |||
| 361 | return !!a; | |||
| 362 | } | |||
| 363 | ||||
| 364 | /* Compute a vector, indexed by character code, of the trie nodes | |||
| 365 | referenced from the given tree. */ | |||
| 366 | static void | |||
| 367 | treenext (struct tree const *tree, struct trie *next[]) | |||
| 368 | { | |||
| 369 | if (!tree) | |||
| 370 | return; | |||
| 371 | treenext(tree->llink, next); | |||
| 372 | treenext(tree->rlink, next); | |||
| 373 | next[tree->label] = tree->trie; | |||
| 374 | } | |||
| 375 | ||||
| 376 | /* Compute the shift for each trie node, as well as the delta | |||
| 377 | table and next cache for the given keyword set. */ | |||
| 378 | const char * | |||
| 379 | kwsprep (kwset_t kws) | |||
| 380 | { | |||
| 381 | register struct kwset *kwset; | |||
| 382 | register int i; | |||
| 383 | register struct trie *curr; | |||
| 384 | register unsigned char const *trans; | |||
| 385 | unsigned char delta[NCHAR((127*2 +1) + 1)]; | |||
| 386 | ||||
| 387 | kwset = (struct kwset *) kws; | |||
| 388 | ||||
| 389 | /* Initial values for the delta table; will be changed later. The | |||
| 390 | delta entry for a given character is the smallest depth of any | |||
| 391 | node at which an outgoing edge is labeled by that character. */ | |||
| 392 | memset(delta, kwset->mind < UCHAR_MAX ? kwset->mind : UCHAR_MAX, NCHAR)__builtin___memset_chk (delta, kwset->mind < (127*2 +1) ? kwset->mind : (127*2 +1), ((127*2 +1) + 1), __builtin_object_size (delta, 0)); | |||
| 393 | ||||
| 394 | /* Check if we can use the simple boyer-moore algorithm, instead | |||
| 395 | of the hairy commentz-walter algorithm. */ | |||
| 396 | if (kwset->words == 1 && kwset->trans == NULL((void*)0)) | |||
| 397 | { | |||
| 398 | char c; | |||
| 399 | ||||
| 400 | /* Looking for just one string. Extract it from the trie. */ | |||
| 401 | kwset->target = obstack_alloc(&kwset->obstack, kwset->mind)__extension__ ({ struct obstack *__h = (&kwset->obstack ); __extension__ ({ struct obstack *__o = (__h); int __len = ( (kwset->mind)); if (__o->chunk_limit - __o->next_free < __len) _obstack_newchunk (__o, __len); ((__o)->next_free += (__len)); (void) 0; }); __extension__ ({ struct obstack * __o1 = (__h); void *__value = (void *) __o1->object_base; if (__o1->next_free == __value) __o1->maybe_empty_object = 1; __o1->next_free = ((sizeof (long int) < sizeof (void *) ? (__o1->object_base) : (char *) 0) + (((__o1->next_free ) - (sizeof (long int) < sizeof (void *) ? (__o1->object_base ) : (char *) 0) + (__o1->alignment_mask)) & ~(__o1-> alignment_mask))); if (__o1->next_free - (char *)__o1-> chunk > __o1->chunk_limit - (char *)__o1->chunk) __o1 ->next_free = __o1->chunk_limit; __o1->object_base = __o1->next_free; __value; }); }); | |||
| 402 | if (!kwset->target) | |||
| 403 | return "memory exhausted"; | |||
| 404 | for (i = kwset->mind - 1, curr = kwset->trie; i >= 0; --i) | |||
| 405 | { | |||
| 406 | kwset->target[i] = curr->links->label; | |||
| 407 | curr = curr->links->trie; | |||
| 408 | } | |||
| 409 | /* Build the Boyer Moore delta. Boy that's easy compared to CW. */ | |||
| 410 | for (i = 0; i < kwset->mind; ++i) | |||
| 411 | delta[U(kwset->target[i])((unsigned char) (kwset->target[i]))] = kwset->mind - (i + 1); | |||
| 412 | /* Find the minimal delta2 shift that we might make after | |||
| 413 | a backwards match has failed. */ | |||
| 414 | c = kwset->target[kwset->mind - 1]; | |||
| 415 | for (i = kwset->mind - 2; i >= 0; --i) | |||
| 416 | if (kwset->target[i] == c) | |||
| 417 | break; | |||
| 418 | kwset->mind2 = kwset->mind - (i + 1); | |||
| 419 | } | |||
| 420 | else | |||
| 421 | { | |||
| 422 | register struct trie *fail; | |||
| 423 | struct trie *last, *next[NCHAR((127*2 +1) + 1)]; | |||
| 424 | ||||
| 425 | /* Traverse the nodes of the trie in level order, simultaneously | |||
| 426 | computing the delta table, failure function, and shift function. */ | |||
| 427 | for (curr = last = kwset->trie; curr; curr = curr->next) | |||
| ||||
| 428 | { | |||
| 429 | /* Enqueue the immediate descendants in the level order queue. */ | |||
| 430 | enqueue(curr->links, &last); | |||
| 431 | ||||
| 432 | curr->shift = kwset->mind; | |||
| 433 | curr->maxshift = kwset->mind; | |||
| 434 | ||||
| 435 | /* Update the delta table for the descendants of this node. */ | |||
| 436 | treedelta(curr->links, curr->depth, delta); | |||
| 437 | ||||
| 438 | /* Compute the failure function for the descendants of this node. */ | |||
| 439 | treefails(curr->links, curr->fail, kwset->trie); | |||
| 440 | ||||
| 441 | /* Update the shifts at each node in the current node's chain | |||
| 442 | of fails back to the root. */ | |||
| 443 | for (fail = curr->fail; fail; fail = fail->fail) | |||
| 444 | { | |||
| 445 | /* If the current node has some outgoing edge that the fail | |||
| 446 | doesn't, then the shift at the fail should be no larger | |||
| 447 | than the difference of their depths. */ | |||
| 448 | if (!hasevery(fail->links, curr->links)) | |||
| 449 | if (curr->depth - fail->depth < fail->shift) | |||
| 450 | fail->shift = curr->depth - fail->depth; | |||
| 451 | ||||
| 452 | /* If the current node is accepting then the shift at the | |||
| 453 | fail and its descendants should be no larger than the | |||
| 454 | difference of their depths. */ | |||
| 455 | if (curr->accepting && fail->maxshift > curr->depth - fail->depth) | |||
| 456 | fail->maxshift = curr->depth - fail->depth; | |||
| 457 | } | |||
| 458 | } | |||
| 459 | ||||
| 460 | /* Traverse the trie in level order again, fixing up all nodes whose | |||
| 461 | shift exceeds their inherited maxshift. */ | |||
| 462 | for (curr = kwset->trie->next; curr; curr = curr->next) | |||
| ||||
| 463 | { | |||
| 464 | if (curr->maxshift > curr->parent->maxshift) | |||
| 465 | curr->maxshift = curr->parent->maxshift; | |||
| 466 | if (curr->shift > curr->maxshift) | |||
| 467 | curr->shift = curr->maxshift; | |||
| 468 | } | |||
| 469 | ||||
| 470 | /* Create a vector, indexed by character code, of the outgoing links | |||
| 471 | from the root node. */ | |||
| 472 | for (i = 0; i < NCHAR((127*2 +1) + 1); ++i) | |||
| 473 | next[i] = NULL((void*)0); | |||
| 474 | treenext(kwset->trie->links, next); | |||
| 475 | ||||
| 476 | if ((trans = kwset->trans) != NULL((void*)0)) | |||
| 477 | for (i = 0; i < NCHAR((127*2 +1) + 1); ++i) | |||
| 478 | kwset->next[i] = next[U(trans[i])((unsigned char) (trans[i]))]; | |||
| 479 | else | |||
| 480 | memcpy(kwset->next, next, NCHAR * sizeof(struct trie *))__builtin___memcpy_chk (kwset->next, next, ((127*2 +1) + 1 ) * sizeof(struct trie *), __builtin_object_size (kwset->next , 0)); | |||
| 481 | } | |||
| 482 | ||||
| 483 | /* Fix things up for any translation table. */ | |||
| 484 | if ((trans = kwset->trans) != NULL((void*)0)) | |||
| 485 | for (i = 0; i < NCHAR((127*2 +1) + 1); ++i) | |||
| 486 | kwset->delta[i] = delta[U(trans[i])((unsigned char) (trans[i]))]; | |||
| 487 | else | |||
| 488 | memcpy(kwset->delta, delta, NCHAR)__builtin___memcpy_chk (kwset->delta, delta, ((127*2 +1) + 1), __builtin_object_size (kwset->delta, 0)); | |||
| 489 | ||||
| 490 | return NULL((void*)0); | |||
| 491 | } | |||
| 492 | ||||
| 493 | /* Fast boyer-moore search. */ | |||
| 494 | static size_t | |||
| 495 | bmexec (kwset_t kws, char const *text, size_t size) | |||
| 496 | { | |||
| 497 | struct kwset const *kwset; | |||
| 498 | register unsigned char const *d1; | |||
| 499 | register char const *ep, *sp, *tp; | |||
| 500 | register int d, gc, i, len, md2; | |||
| 501 | ||||
| 502 | kwset = (struct kwset const *) kws; | |||
| 503 | len = kwset->mind; | |||
| 504 | ||||
| 505 | if (len == 0) | |||
| 506 | return 0; | |||
| 507 | if (len > size) | |||
| 508 | return -1; | |||
| 509 | if (len == 1) | |||
| 510 | { | |||
| 511 | tp = memchr (text, kwset->target[0], size); | |||
| 512 | return tp ? tp - text : -1; | |||
| 513 | } | |||
| 514 | ||||
| 515 | d1 = kwset->delta; | |||
| 516 | sp = kwset->target + len; | |||
| 517 | gc = U(sp[-2])((unsigned char) (sp[-2])); | |||
| 518 | md2 = kwset->mind2; | |||
| 519 | tp = text + len; | |||
| 520 | ||||
| 521 | /* Significance of 12: 1 (initial offset) + 10 (skip loop) + 1 (md2). */ | |||
| 522 | if (size > 12 * len) | |||
| 523 | /* 11 is not a bug, the initial offset happens only once. */ | |||
| 524 | for (ep = text + size - 11 * len;;) | |||
| 525 | { | |||
| 526 | while (tp <= ep) | |||
| 527 | { | |||
| 528 | d = d1[U(tp[-1])((unsigned char) (tp[-1]))], tp += d; | |||
| 529 | d = d1[U(tp[-1])((unsigned char) (tp[-1]))], tp += d; | |||
| 530 | if (d == 0) | |||
| 531 | goto found; | |||
| 532 | d = d1[U(tp[-1])((unsigned char) (tp[-1]))], tp += d; | |||
| 533 | d = d1[U(tp[-1])((unsigned char) (tp[-1]))], tp += d; | |||
| 534 | d = d1[U(tp[-1])((unsigned char) (tp[-1]))], tp += d; | |||
| 535 | if (d == 0) | |||
| 536 | goto found; | |||
| 537 | d = d1[U(tp[-1])((unsigned char) (tp[-1]))], tp += d; | |||
| 538 | d = d1[U(tp[-1])((unsigned char) (tp[-1]))], tp += d; | |||
| 539 | d = d1[U(tp[-1])((unsigned char) (tp[-1]))], tp += d; | |||
| 540 | if (d == 0) | |||
| 541 | goto found; | |||
| 542 | d = d1[U(tp[-1])((unsigned char) (tp[-1]))], tp += d; | |||
| 543 | d = d1[U(tp[-1])((unsigned char) (tp[-1]))], tp += d; | |||
| 544 | } | |||
| 545 | break; | |||
| 546 | found: | |||
| 547 | if (U(tp[-2])((unsigned char) (tp[-2])) == gc) | |||
| 548 | { | |||
| 549 | for (i = 3; i <= len && U(tp[-i])((unsigned char) (tp[-i])) == U(sp[-i])((unsigned char) (sp[-i])); ++i) | |||
| 550 | ; | |||
| 551 | if (i > len) | |||
| 552 | return tp - len - text; | |||
| 553 | } | |||
| 554 | tp += md2; | |||
| 555 | } | |||
| 556 | ||||
| 557 | /* Now we have only a few characters left to search. We | |||
| 558 | carefully avoid ever producing an out-of-bounds pointer. */ | |||
| 559 | ep = text + size; | |||
| 560 | d = d1[U(tp[-1])((unsigned char) (tp[-1]))]; | |||
| 561 | while (d <= ep - tp) | |||
| 562 | { | |||
| 563 | d = d1[U((tp += d)[-1])((unsigned char) ((tp += d)[-1]))]; | |||
| 564 | if (d != 0) | |||
| 565 | continue; | |||
| 566 | if (U(tp[-2])((unsigned char) (tp[-2])) == gc) | |||
| 567 | { | |||
| 568 | for (i = 3; i <= len && U(tp[-i])((unsigned char) (tp[-i])) == U(sp[-i])((unsigned char) (sp[-i])); ++i) | |||
| 569 | ; | |||
| 570 | if (i > len) | |||
| 571 | return tp - len - text; | |||
| 572 | } | |||
| 573 | d = md2; | |||
| 574 | } | |||
| 575 | ||||
| 576 | return -1; | |||
| 577 | } | |||
| 578 | ||||
| 579 | /* Hairy multiple string search. */ | |||
| 580 | static size_t | |||
| 581 | cwexec (kwset_t kws, char const *text, size_t len, struct kwsmatch *kwsmatch) | |||
| 582 | { | |||
| 583 | struct kwset const *kwset; | |||
| 584 | struct trie * const *next; | |||
| 585 | struct trie const *trie; | |||
| 586 | struct trie const *accept; | |||
| 587 | char const *beg, *lim, *mch, *lmch; | |||
| 588 | register unsigned char c; | |||
| 589 | register unsigned char const *delta; | |||
| 590 | register int d; | |||
| 591 | register char const *end, *qlim; | |||
| 592 | register struct tree const *tree; | |||
| 593 | register unsigned char const *trans; | |||
| 594 | ||||
| 595 | accept = NULL((void*)0); | |||
| 596 | ||||
| 597 | /* Initialize register copies and look for easy ways out. */ | |||
| 598 | kwset = (struct kwset *) kws; | |||
| 599 | if (len < kwset->mind) | |||
| 600 | return -1; | |||
| 601 | next = kwset->next; | |||
| 602 | delta = kwset->delta; | |||
| 603 | trans = kwset->trans; | |||
| 604 | lim = text + len; | |||
| 605 | end = text; | |||
| 606 | if ((d = kwset->mind) != 0) | |||
| 607 | mch = NULL((void*)0); | |||
| 608 | else | |||
| 609 | { | |||
| 610 | mch = text, accept = kwset->trie; | |||
| 611 | goto match; | |||
| 612 | } | |||
| 613 | ||||
| 614 | if (len >= 4 * kwset->mind) | |||
| 615 | qlim = lim - 4 * kwset->mind; | |||
| 616 | else | |||
| 617 | qlim = NULL((void*)0); | |||
| 618 | ||||
| 619 | while (lim - end >= d) | |||
| 620 | { | |||
| 621 | if (qlim && end <= qlim) | |||
| 622 | { | |||
| 623 | end += d - 1; | |||
| 624 | while ((d = delta[c = *end]) && end < qlim) | |||
| 625 | { | |||
| 626 | end += d; | |||
| 627 | end += delta[U(*end)((unsigned char) (*end))]; | |||
| 628 | end += delta[U(*end)((unsigned char) (*end))]; | |||
| 629 | } | |||
| 630 | ++end; | |||
| 631 | } | |||
| 632 | else | |||
| 633 | d = delta[c = (end += d)[-1]]; | |||
| 634 | if (d) | |||
| 635 | continue; | |||
| 636 | beg = end - 1; | |||
| 637 | trie = next[c]; | |||
| 638 | if (trie->accepting) | |||
| 639 | { | |||
| 640 | mch = beg; | |||
| 641 | accept = trie; | |||
| 642 | } | |||
| 643 | d = trie->shift; | |||
| 644 | while (beg > text) | |||
| 645 | { | |||
| 646 | c = trans ? trans[U(*--beg)((unsigned char) (*--beg))] : *--beg; | |||
| 647 | tree = trie->links; | |||
| 648 | while (tree && c != tree->label) | |||
| 649 | if (c < tree->label) | |||
| 650 | tree = tree->llink; | |||
| 651 | else | |||
| 652 | tree = tree->rlink; | |||
| 653 | if (tree) | |||
| 654 | { | |||
| 655 | trie = tree->trie; | |||
| 656 | if (trie->accepting) | |||
| 657 | { | |||
| 658 | mch = beg; | |||
| 659 | accept = trie; | |||
| 660 | } | |||
| 661 | } | |||
| 662 | else | |||
| 663 | break; | |||
| 664 | d = trie->shift; | |||
| 665 | } | |||
| 666 | if (mch) | |||
| 667 | goto match; | |||
| 668 | } | |||
| 669 | return -1; | |||
| 670 | ||||
| 671 | match: | |||
| 672 | /* Given a known match, find the longest possible match anchored | |||
| 673 | at or before its starting point. This is nearly a verbatim | |||
| 674 | copy of the preceding main search loops. */ | |||
| 675 | if (lim - mch > kwset->maxd) | |||
| 676 | lim = mch + kwset->maxd; | |||
| 677 | lmch = NULL((void*)0); | |||
| 678 | d = 1; | |||
| 679 | while (lim - end >= d) | |||
| 680 | { | |||
| 681 | if ((d = delta[c = (end += d)[-1]]) != 0) | |||
| 682 | continue; | |||
| 683 | beg = end - 1; | |||
| 684 | if (!(trie = next[c])) | |||
| 685 | { | |||
| 686 | d = 1; | |||
| 687 | continue; | |||
| 688 | } | |||
| 689 | if (trie->accepting && beg <= mch) | |||
| 690 | { | |||
| 691 | lmch = beg; | |||
| 692 | accept = trie; | |||
| 693 | } | |||
| 694 | d = trie->shift; | |||
| 695 | while (beg > text) | |||
| 696 | { | |||
| 697 | c = trans ? trans[U(*--beg)((unsigned char) (*--beg))] : *--beg; | |||
| 698 | tree = trie->links; | |||
| 699 | while (tree && c != tree->label) | |||
| 700 | if (c < tree->label) | |||
| 701 | tree = tree->llink; | |||
| 702 | else | |||
| 703 | tree = tree->rlink; | |||
| 704 | if (tree) | |||
| 705 | { | |||
| 706 | trie = tree->trie; | |||
| 707 | if (trie->accepting && beg <= mch) | |||
| 708 | { | |||
| 709 | lmch = beg; | |||
| 710 | accept = trie; | |||
| 711 | } | |||
| 712 | } | |||
| 713 | else | |||
| 714 | break; | |||
| 715 | d = trie->shift; | |||
| 716 | } | |||
| 717 | if (lmch) | |||
| 718 | { | |||
| 719 | mch = lmch; | |||
| 720 | goto match; | |||
| 721 | } | |||
| 722 | if (!d) | |||
| 723 | d = 1; | |||
| 724 | } | |||
| 725 | ||||
| 726 | if (kwsmatch) | |||
| 727 | { | |||
| 728 | kwsmatch->index = accept->accepting / 2; | |||
| 729 | kwsmatch->offset[0] = mch - text; | |||
| 730 | kwsmatch->size[0] = accept->depth; | |||
| 731 | } | |||
| 732 | return mch - text; | |||
| 733 | } | |||
| 734 | ||||
| 735 | /* Search through the given text for a match of any member of the | |||
| 736 | given keyword set. Return a pointer to the first character of | |||
| 737 | the matching substring, or NULL if no match is found. If FOUNDLEN | |||
| 738 | is non-NULL store in the referenced location the length of the | |||
| 739 | matching substring. Similarly, if FOUNDIDX is non-NULL, store | |||
| 740 | in the referenced location the index number of the particular | |||
| 741 | keyword matched. */ | |||
| 742 | size_t | |||
| 743 | kwsexec (kwset_t kws, char const *text, size_t size, | |||
| 744 | struct kwsmatch *kwsmatch) | |||
| 745 | { | |||
| 746 | struct kwset const *kwset = (struct kwset *) kws; | |||
| 747 | if (kwset->words == 1 && kwset->trans == NULL((void*)0)) | |||
| 748 | { | |||
| 749 | size_t ret = bmexec (kws, text, size); | |||
| 750 | if (kwsmatch != NULL((void*)0) && ret != (size_t) -1) | |||
| 751 | { | |||
| 752 | kwsmatch->index = 0; | |||
| 753 | kwsmatch->offset[0] = ret; | |||
| 754 | kwsmatch->size[0] = kwset->mind; | |||
| 755 | } | |||
| 756 | return ret; | |||
| 757 | } | |||
| 758 | else | |||
| 759 | return cwexec(kws, text, size, kwsmatch); | |||
| 760 | } | |||
| 761 | ||||
| 762 | /* Free the components of the given keyword set. */ | |||
| 763 | void | |||
| 764 | kwsfree (kwset_t kws) | |||
| 765 | { | |||
| 766 | struct kwset *kwset; | |||
| 767 | ||||
| 768 | kwset = (struct kwset *) kws; | |||
| 769 | obstack_free(&kwset->obstack, NULL)__extension__ ({ struct obstack *__o = (&kwset->obstack ); void *__obj = (((void*)0)); if (__obj > (void *)__o-> chunk && __obj < (void *)__o->chunk_limit) __o-> next_free = __o->object_base = (char *)__obj; else (obstack_free ) (__o, __obj); }); | |||
| 770 | free(kws); | |||
| 771 | } |