1 /* Host name resolution and matching.
2 Copyright (C) 1995, 1996, 1997, 2000, 2001 Free Software Foundation, Inc.
4 This file is part of GNU Wget.
6 GNU Wget is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2 of the License, or
9 (at your option) any later version.
11 GNU Wget 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
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with Wget; if not, write to the Free Software
18 Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */
30 #include <sys/types.h>
35 # include <sys/socket.h>
36 # include <netinet/in.h>
38 # include <arpa/inet.h>
44 #define NO_ADDRESS NO_DATA
47 #ifdef HAVE_SYS_UTSNAME_H
48 # include <sys/utsname.h>
68 #define IP4_ADDRESS_LENGTH 4
70 /* Mapping between known hosts and to lists of their addresses. */
72 static struct hash_table *host_name_addresses_map;
74 /* Lists of addresses. This should eventually be extended to handle
78 int count; /* number of adrresses */
79 unsigned char *buffer; /* buffer which holds all of them. */
81 int faulty; /* number of addresses known not to
83 int refcount; /* so we know whether to free it or
87 #define ADDR_LOCATION(al, index) ((al)->buffer + index * IP4_ADDRESS_LENGTH)
89 /* Get the bounds of the address list. */
92 address_list_get_bounds (struct address_list *al, int *start, int *end)
98 /* Copy address number INDEX to IP_STORE. */
101 address_list_copy_one (struct address_list *al, int index,
102 unsigned char *ip_store)
104 assert (index >= al->faulty && index < al->count);
105 memcpy (ip_store, ADDR_LOCATION (al, index), IP4_ADDRESS_LENGTH);
108 /* Check whether two address lists have all their IPs in common. */
111 address_list_match_all (struct address_list *al1, struct address_list *al2)
115 if (al1->count != al2->count)
117 return 0 == memcmp (al1->buffer, al2->buffer,
118 al1->count * IP4_ADDRESS_LENGTH);
121 /* Mark the INDEXth element of AL as faulty, so that the next time
122 this address list is used, the faulty element will be skipped. */
125 address_list_set_faulty (struct address_list *al, int index)
127 /* We assume that the address list is traversed in order, so that a
128 "faulty" attempt is always preceded with all-faulty addresses,
129 and this is how Wget uses it. */
130 assert (index == al->faulty);
133 if (al->faulty >= al->count)
134 /* All addresses have been proven faulty. Since there's not much
135 sense in returning the user an empty address list the next
136 time, we'll rather make them all clean, so that they can be
141 /* Create an address_list out of a NULL-terminated list of addresses,
142 as returned by gethostbyname. */
144 static struct address_list *
145 address_list_new (char **h_addr_list)
149 struct address_list *al = xmalloc (sizeof (struct address_list));
151 while (h_addr_list[count])
156 al->buffer = xmalloc (count * IP4_ADDRESS_LENGTH);
159 for (i = 0; i < count; i++)
160 memcpy (ADDR_LOCATION (al, i), h_addr_list[i], IP4_ADDRESS_LENGTH);
165 /* Like address_list_new, but initialized with only one address. */
167 static struct address_list *
168 address_list_new_one (const char *addr)
170 struct address_list *al = xmalloc (sizeof (struct address_list));
173 al->buffer = xmalloc (IP4_ADDRESS_LENGTH);
175 memcpy (ADDR_LOCATION (al, 0), addr, IP4_ADDRESS_LENGTH);
181 address_list_delete (struct address_list *al)
188 address_list_release (struct address_list *al)
191 DEBUGP (("Releasing %p (new refcount %d).\n", al, al->refcount));
192 if (al->refcount <= 0)
194 DEBUGP (("Deleting unused %p.\n", al));
195 address_list_delete (al);
199 /* The same as inet_ntoa, but without the need for a cast, or for
200 #including the netinet stuff. */
203 pretty_print_address (const unsigned char *addr)
205 return inet_ntoa (*(struct in_addr *)addr);
208 /* Add host name HOST with the address ADDR_TEXT to the cache.
209 ADDR_LIST is a NULL-terminated list of addresses, as in struct
213 cache_host_lookup (const char *host, struct address_list *al)
215 if (!host_name_addresses_map)
216 host_name_addresses_map = make_nocase_string_hash_table (0);
219 hash_table_put (host_name_addresses_map, xstrdup_lower (host), al);
225 debug_logprintf ("Caching %s =>", host);
226 for (i = 0; i < al->count; i++)
227 debug_logprintf (" %s",
228 pretty_print_address (ADDR_LOCATION (al, i)));
229 debug_logprintf ("\n");
234 struct address_list *
235 lookup_host (const char *host, int silent)
237 struct address_list *al = NULL;
239 struct hostent *hptr;
241 /* If the address is of the form d.d.d.d, no further lookup is
243 addr = (unsigned long)inet_addr (host);
246 /* ADDR is defined to be in network byte order, which is what
247 this returns, so we can just copy it to STORE_IP. However,
248 on big endian 64-bit architectures the value will be stored
249 in the *last*, not first four bytes. OFFSET makes sure that
250 we copy the correct four bytes. */
252 #ifdef WORDS_BIGENDIAN
253 offset = sizeof (unsigned long) - IP4_ADDRESS_LENGTH;
257 return address_list_new_one ((char *)&addr + offset);
260 /* By now we know that the host name we got is not of the form
261 d.d.d.d. Try to find it in our cache of host names. */
262 if (host_name_addresses_map)
263 al = hash_table_get (host_name_addresses_map, host);
267 DEBUGP (("Found %s in host_name_addresses_map (%p)\n", host, al));
273 logprintf (LOG_VERBOSE, _("Resolving %s... "), host);
275 /* Look up the host using gethostbyname(). */
276 hptr = gethostbyname (host);
280 logprintf (LOG_VERBOSE, _("failed: %s.\n"), herrmsg (h_errno));
285 logprintf (LOG_VERBOSE, _("done.\n"));
287 /* Do all systems have h_addr_list, or is it a newer thing? If the
288 latter, use address_list_new_one. */
289 al = address_list_new (hptr->h_addr_list);
291 /* Cache the lookup information. */
292 cache_host_lookup (host, al);
297 /* Determine whether a URL is acceptable to be followed, according to
298 a list of domains to accept. */
300 accept_domain (struct url *u)
302 assert (u->host != NULL);
305 if (!sufmatch ((const char **)opt.domains, u->host))
308 if (opt.exclude_domains)
310 if (sufmatch ((const char **)opt.exclude_domains, u->host))
316 /* Check whether WHAT is matched in LIST, each element of LIST being a
317 pattern to match WHAT against, using backward matching (see
318 match_backwards() in utils.c).
320 If an element of LIST matched, 1 is returned, 0 otherwise. */
322 sufmatch (const char **list, const char *what)
327 for (i = 0; list[i]; i++)
329 for (j = strlen (list[i]), k = lw; j >= 0 && k >= 0; j--, k--)
330 if (TOLOWER (list[i][j]) != TOLOWER (what[k]))
332 /* The domain must be first to reach to beginning. */
339 /* Print error messages for host errors. */
343 /* Can't use switch since some constants are equal (at least on my
344 system), and the compiler signals "duplicate case value". */
345 if (error == HOST_NOT_FOUND
346 || error == NO_RECOVERY
348 || error == NO_ADDRESS
349 || error == TRY_AGAIN)
350 return _("Host not found");
352 return _("Unknown error");
356 host_cleanup_mapper (void *key, void *value, void *arg_ignored)
358 struct address_list *al;
360 xfree (key); /* host */
362 al = (struct address_list *)value;
363 assert (al->refcount == 1);
364 address_list_delete (al);
372 if (host_name_addresses_map)
374 hash_table_map (host_name_addresses_map, host_cleanup_mapper, NULL);
375 hash_table_destroy (host_name_addresses_map);
376 host_name_addresses_map = NULL;