]>
Commit | Line | Data |
---|---|---|
14c63c25 | 1 | |
2 | /* | |
262a0e14 | 3 | * $Id$ |
14c63c25 | 4 | * |
5 | * DEBUG: section 0 Hash Tables | |
6 | * AUTHOR: Harvest Derived | |
7 | * | |
2b6662ba | 8 | * SQUID Web Proxy Cache http://www.squid-cache.org/ |
9 | * ---------------------------------------------------------- | |
14c63c25 | 10 | * |
2b6662ba | 11 | * Squid is the result of efforts by numerous individuals from |
12 | * the Internet community; see the CONTRIBUTORS file for full | |
13 | * details. Many organizations have provided support for Squid's | |
14 | * development; see the SPONSORS file for full details. Squid is | |
15 | * Copyrighted (C) 2001 by the Regents of the University of | |
16 | * California; see the COPYRIGHT file for full details. Squid | |
17 | * incorporates software developed and/or copyrighted by other | |
18 | * sources; see the CREDITS file for full details. | |
14c63c25 | 19 | * |
20 | * This program is free software; you can redistribute it and/or modify | |
21 | * it under the terms of the GNU General Public License as published by | |
22 | * the Free Software Foundation; either version 2 of the License, or | |
23 | * (at your option) any later version. | |
26ac0430 | 24 | * |
14c63c25 | 25 | * This program is distributed in the hope that it will be useful, |
26 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
27 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
28 | * GNU General Public License for more details. | |
26ac0430 | 29 | * |
14c63c25 | 30 | * You should have received a copy of the GNU General Public License |
31 | * along with this program; if not, write to the Free Software | |
cbdec147 | 32 | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA. |
26ac0430 | 33 | * |
14c63c25 | 34 | */ |
35 | ||
14c63c25 | 36 | #include <unistd.h> |
37 | #include <stdlib.h> | |
38 | #include <stdio.h> | |
39 | #include <sys/types.h> | |
40 | #include <ctype.h> | |
41 | #include <sys/time.h> | |
42 | #include <strings.h> | |
e7740002 | 43 | #include <assert.h> |
14c63c25 | 44 | #include "hash.h" |
e7740002 | 45 | #undef free |
468ae12b | 46 | extern void my_free(char *, int, void *); |
e7740002 | 47 | |
48 | #define free(a) my_free(__FILE__, __LINE__, a) | |
14c63c25 | 49 | |
da537f52 | 50 | extern void print_stats(); |
14c63c25 | 51 | /* |
52 | * hash_url() - Returns a well-distributed hash function for URLs. | |
53 | * The best way is to sum up the last half of the string. | |
54 | * Adapted from code written by Mic Bowman. -Darren | |
55 | * Generates a standard deviation = 15.73 | |
56 | */ | |
57 | unsigned int | |
58 | hash_url(const void *data, unsigned int size) | |
59 | { | |
60 | const char *s = data; | |
61 | unsigned int i, j, n; | |
62 | j = strlen(s); | |
63 | for (i = j / 2, n = 0; i < j; i++) | |
26ac0430 | 64 | n ^= 271 * (unsigned) s[i]; |
14c63c25 | 65 | i = n ^ (j * 271); |
66 | return i % size; | |
67 | } | |
68 | ||
69 | unsigned int | |
70 | hash_string(const void *data, unsigned int size) | |
71 | { | |
72 | const char *s = data; | |
73 | unsigned int n = 0; | |
74 | unsigned int j = 0; | |
75 | unsigned int i = 0; | |
76 | while (*s) { | |
26ac0430 AJ |
77 | j++; |
78 | n ^= 271 * (unsigned) *s++; | |
14c63c25 | 79 | } |
80 | i = n ^ (j * 271); | |
81 | return i % size; | |
82 | } | |
83 | ||
84 | /* the following 4 functions were adapted from | |
85 | * usr/src/lib/libc/db/hash_func.c, 4.4 BSD lite */ | |
86 | ||
87 | /* | |
88 | * HASH FUNCTIONS | |
89 | * | |
90 | * Assume that we've already split the bucket to which this key hashes, | |
91 | * calculate that bucket, and check that in fact we did already split it. | |
92 | * | |
93 | * This came from ejb's hsearch. | |
94 | */ | |
95 | ||
96 | #define PRIME1 37 | |
97 | #define PRIME2 1048583 | |
98 | ||
99 | /* Hash function from Chris Torek. */ | |
100 | unsigned int | |
101 | hash4(const void *data, unsigned int size) | |
102 | { | |
103 | const char *key = data; | |
104 | size_t loop; | |
105 | unsigned int h; | |
106 | size_t len; | |
107 | ||
108 | #define HASH4a h = (h << 5) - h + *key++; | |
109 | #define HASH4b h = (h << 5) + h + *key++; | |
110 | #define HASH4 HASH4b | |
111 | ||
112 | h = 0; | |
113 | len = strlen(key); | |
114 | loop = (len + 8 - 1) >> 3; | |
115 | switch (len & (8 - 1)) { | |
116 | case 0: | |
26ac0430 AJ |
117 | do { |
118 | HASH4; | |
119 | /* FALLTHROUGH */ | |
120 | case 7: | |
121 | HASH4; | |
122 | /* FALLTHROUGH */ | |
123 | case 6: | |
124 | HASH4; | |
125 | /* FALLTHROUGH */ | |
126 | case 5: | |
127 | HASH4; | |
128 | /* FALLTHROUGH */ | |
129 | case 4: | |
130 | HASH4; | |
131 | /* FALLTHROUGH */ | |
132 | case 3: | |
133 | HASH4; | |
134 | /* FALLTHROUGH */ | |
135 | case 2: | |
136 | HASH4; | |
137 | /* FALLTHROUGH */ | |
138 | case 1: | |
139 | HASH4; | |
140 | } while (--loop); | |
14c63c25 | 141 | } |
142 | return h % size; | |
143 | } | |
144 | ||
145 | /* | |
146 | * hash_create - creates a new hash table, uses the cmp_func | |
147 | * to compare keys. Returns the identification for the hash table; | |
148 | * otherwise returns a negative number on error. | |
149 | */ | |
150 | hash_table * | |
151 | hash_create(HASHCMP * cmp_func, int hash_sz, HASHHASH * hash_func) | |
152 | { | |
153 | hash_table *hid = calloc(1, sizeof(hash_table)); | |
154 | if (!hash_sz) | |
26ac0430 | 155 | hid->size = (unsigned int) DEFAULT_HASH_SIZE; |
14c63c25 | 156 | else |
26ac0430 | 157 | hid->size = (unsigned int) hash_sz; |
14c63c25 | 158 | /* allocate and null the buckets */ |
159 | hid->buckets = calloc(hid->size, sizeof(hash_link *)); | |
160 | hid->cmp = cmp_func; | |
161 | hid->hash = hash_func; | |
162 | hid->current_ptr = NULL; | |
163 | hid->current_slot = 0; | |
164 | return hid; | |
165 | } | |
166 | ||
167 | /* | |
168 | * hash_insert - inserts the given item 'item' under the given key 'k' | |
169 | * into the hash table 'hid'. Returns non-zero on error; otherwise, | |
170 | * returns 0 and inserts the item. | |
171 | * | |
172 | * It does not copy any data into the hash table, only pointers. | |
173 | */ | |
174 | void | |
175 | hash_insert(hash_table * hid, const char *k, void *item) | |
176 | { | |
177 | int i; | |
178 | hash_link *new; | |
179 | assert(k != NULL); | |
180 | /* Add to the given hash table 'hid' */ | |
181 | new = calloc(1, sizeof(hash_link)); | |
da537f52 | 182 | if (!new) { |
26ac0430 AJ |
183 | fprintf(stderr, "calloc failed!\n"); |
184 | print_stats(); | |
185 | exit(1); | |
da537f52 | 186 | } |
14c63c25 | 187 | new->item = item; |
188 | new->key = (char *) k; | |
189 | i = hid->hash(k, hid->size); | |
190 | new->next = hid->buckets[i]; | |
191 | hid->buckets[i] = new; | |
192 | } | |
193 | ||
194 | /* | |
195 | * hash_join - joins a hash_link under its key lnk->key | |
26ac0430 | 196 | * into the hash table 'hid'. |
14c63c25 | 197 | * |
198 | * It does not copy any data into the hash table, only links pointers. | |
199 | */ | |
200 | void | |
201 | hash_join(hash_table * hid, hash_link * lnk) | |
202 | { | |
203 | int i; | |
204 | i = hid->hash(lnk->key, hid->size); | |
205 | lnk->next = hid->buckets[i]; | |
206 | hid->buckets[i] = lnk; | |
207 | } | |
208 | ||
209 | /* | |
210 | * hash_lookup - locates the item under the key 'k' in the hash table | |
211 | * 'hid'. Returns a pointer to the hash bucket on success; otherwise | |
212 | * returns NULL. | |
213 | */ | |
214 | hash_link * | |
215 | hash_lookup(hash_table * hid, const void *k) | |
216 | { | |
217 | hash_link *walker; | |
218 | int b; | |
219 | assert(k != NULL); | |
220 | b = hid->hash(k, hid->size); | |
221 | for (walker = hid->buckets[b]; walker != NULL; walker = walker->next) { | |
26ac0430 AJ |
222 | if ((hid->cmp) (k, walker->key) == 0) |
223 | return (walker); | |
224 | assert(walker != walker->next); | |
14c63c25 | 225 | } |
226 | return NULL; | |
227 | } | |
228 | ||
229 | /* | |
230 | * hash_first - returns the first item in the hash table 'hid'. | |
231 | * Otherwise, returns NULL on error. | |
232 | */ | |
233 | hash_link * | |
234 | hash_first(hash_table * hid) | |
235 | { | |
236 | int i; | |
237 | ||
238 | for (i = 0; i < hid->size; i++) { | |
26ac0430 AJ |
239 | hid->current_slot = i; |
240 | if (hid->buckets[i] != NULL) | |
241 | return (hid->current_ptr = hid->buckets[i]); | |
14c63c25 | 242 | } |
243 | return NULL; | |
244 | } | |
245 | ||
246 | /* | |
247 | * hash_next - returns the next item in the hash table 'hid'. | |
26ac0430 | 248 | * Otherwise, returns NULL on error or end of list. |
14c63c25 | 249 | * |
250 | * MUST call hash_first() before hash_next(). | |
251 | */ | |
252 | hash_link * | |
253 | hash_next(hash_table * hid) | |
254 | { | |
255 | int i; | |
256 | ||
257 | if (hid->current_ptr != NULL) { | |
26ac0430 AJ |
258 | hid->current_ptr = hid->current_ptr->next; |
259 | if (hid->current_ptr != NULL) | |
260 | return (hid->current_ptr); /* next item */ | |
14c63c25 | 261 | } |
262 | /* find next bucket */ | |
263 | for (i = hid->current_slot + 1; i < hid->size; i++) { | |
26ac0430 AJ |
264 | hid->current_slot = i; |
265 | if (hid->buckets[i] != NULL) | |
266 | return (hid->current_ptr = hid->buckets[i]); | |
14c63c25 | 267 | } |
268 | return NULL; /* end of list */ | |
269 | } | |
270 | ||
271 | int | |
272 | hash_delete(hash_table * hid, const char *key) | |
273 | { | |
274 | return hash_delete_link(hid, hash_lookup(hid, key)); | |
275 | } | |
276 | ||
277 | /* | |
26ac0430 | 278 | * hash_delete_link - deletes the given hash_link node from the |
14c63c25 | 279 | * hash table 'hid'. If FreeLink then free the given hash_link. |
280 | * | |
26ac0430 | 281 | * On success, it returns 0 and deletes the link; otherwise, |
14c63c25 | 282 | * returns non-zero on error. |
283 | */ | |
da537f52 | 284 | int |
14c63c25 | 285 | hash_unlink(hash_table * hid, hash_link * hl, int FreeLink) |
286 | { | |
287 | hash_link *walker, *prev; | |
288 | int i; | |
289 | if (hl == NULL) | |
26ac0430 | 290 | return -1; |
14c63c25 | 291 | i = hid->hash(hl->key, hid->size); |
292 | for (prev = NULL, walker = hid->buckets[i]; | |
26ac0430 AJ |
293 | walker != NULL; prev = walker, walker = walker->next) { |
294 | if (walker == hl) { | |
295 | if (prev == NULL) { /* it's the head */ | |
296 | hid->buckets[i] = walker->next; | |
297 | } else { | |
298 | prev->next = walker->next; /* skip it */ | |
299 | } | |
300 | /* fix walker state if needed */ | |
301 | if (walker == hid->current_ptr) | |
302 | hid->current_ptr = walker->next; | |
303 | if (FreeLink) { | |
304 | if (walker) { | |
305 | free(walker); | |
306 | } | |
307 | } | |
308 | return 0; | |
309 | } | |
14c63c25 | 310 | } |
311 | return 1; | |
312 | } | |
313 | ||
314 | /* take link off and free link node */ | |
315 | int | |
316 | hash_delete_link(hash_table * hid, hash_link * hl) | |
317 | { | |
318 | return (hash_unlink(hid, hl, 1)); | |
319 | } | |
320 | ||
321 | /* take link off only */ | |
322 | int | |
323 | hash_remove_link(hash_table * hid, hash_link * hl) | |
324 | { | |
325 | return (hash_unlink(hid, hl, 0)); | |
326 | } | |
327 | ||
328 | /* | |
26ac0430 | 329 | * hash_get_bucket - returns the head item of the bucket |
14c63c25 | 330 | * in the hash table 'hid'. Otherwise, returns NULL on error. |
331 | */ | |
332 | hash_link * | |
333 | hash_get_bucket(hash_table * hid, unsigned int bucket) | |
334 | { | |
335 | if (bucket >= hid->size) | |
26ac0430 | 336 | return NULL; |
14c63c25 | 337 | return (hid->buckets[bucket]); |
338 | } | |
339 | ||
340 | ||
341 | void | |
342 | hashFreeMemory(hash_table * hid) | |
343 | { | |
468ae12b | 344 | if (hid->buckets); |
14c63c25 | 345 | free(hid->buckets); |
468ae12b | 346 | if (hid) |
26ac0430 | 347 | free(hid); |
14c63c25 | 348 | } |
349 | ||
350 | ||
351 | #ifdef USE_HASH_DRIVER | |
352 | /* | |
353 | * hash-driver - Run with a big file as stdin to insert each line into the | |
354 | * hash table, then prints the whole hash table, then deletes a random item, | |
355 | * and prints the table again... | |
356 | */ | |
357 | int | |
358 | main(void) | |
359 | { | |
360 | hash_table *hid; | |
361 | int i; | |
362 | LOCAL_ARRAY(char, buf, BUFSIZ); | |
363 | LOCAL_ARRAY(char, todelete, BUFSIZ); | |
364 | hash_link *walker = NULL; | |
365 | ||
366 | todelete[0] = '\0'; | |
367 | printf("init\n"); | |
368 | ||
369 | printf("creating hash table\n"); | |
370 | if ((hid = hash_create(strcmp, 229, hash_string)) < 0) { | |
26ac0430 AJ |
371 | printf("hash_create error.\n"); |
372 | exit(1); | |
14c63c25 | 373 | } |
374 | printf("done creating hash table: %d\n", hid); | |
375 | ||
376 | while (fgets(buf, BUFSIZ, stdin)) { | |
26ac0430 AJ |
377 | buf[strlen(buf) - 1] = '\0'; |
378 | printf("Inserting '%s' for item %p to hash table: %d\n", | |
379 | buf, buf, hid); | |
380 | hash_insert(hid, xstrdup(buf), (void *) 0x12345678); | |
381 | if (random() % 17 == 0) | |
382 | strcpy(todelete, buf); | |
14c63c25 | 383 | } |
384 | ||
385 | printf("walking hash table...\n"); | |
386 | for (i = 0, walker = hash_first(hid); walker; walker = hash_next(hid)) { | |
26ac0430 AJ |
387 | printf("item %5d: key: '%s' item: %p\n", i++, walker->key, |
388 | walker->item); | |
14c63c25 | 389 | } |
390 | printf("done walking hash table...\n"); | |
391 | ||
392 | if (todelete[0]) { | |
26ac0430 AJ |
393 | printf("deleting %s from %d\n", todelete, hid); |
394 | if (hash_delete(hid, todelete)) | |
395 | printf("hash_delete error\n"); | |
14c63c25 | 396 | } |
397 | printf("walking hash table...\n"); | |
398 | for (i = 0, walker = hash_first(hid); walker; walker = hash_next(hid)) { | |
26ac0430 AJ |
399 | printf("item %5d: key: '%s' item: %p\n", i++, walker->key, |
400 | walker->item); | |
14c63c25 | 401 | } |
402 | printf("done walking hash table...\n"); | |
403 | ||
404 | ||
405 | printf("driver finished.\n"); | |
406 | exit(0); | |
407 | } | |
408 | #endif |