From 8e330fdba016c5e521ea8e9c11e6df011c8666b1 Mon Sep 17 00:00:00 2001
From: hniksic <devnull@localhost>
Date: Fri, 14 Nov 2003 17:49:52 -0800
Subject: [PATCH] [svn] Add support for -4/-6.
---
src/ChangeLog | 17 ++++++
src/connect.c | 7 ++-
src/host.c | 150 ++++++++++++++++++++++++++++----------------------
src/host.h | 9 +--
src/init.c | 2 +-
5 files changed, 110 insertions(+), 75 deletions(-)
diff --git a/src/ChangeLog b/src/ChangeLog
index 7768b9a0..270d937c 100644
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,20 @@
+2003-11-15 Hrvoje Niksic <hniksic@xemacs.org>
+
+ * init.c: Make sure the options are in alphabetic order!
+
+ * host.c (lookup_host): Merge lookup_host_passive and lookup_host
+ after all -- having both would result in some code duplication.
+ (lookup_host): Set hints.ai_family to AF_INET if ipv4_only is
+ requested. Likewise, set it to AF_INET6 for ipv6_only. Specify
+ AI_ADDRCONFIG where available.
+ (lookup_host): New flag LH_REFRESH that specifies that a cached
+ entry for HOST should be refreshed.
+ (cache_query): New function.
+ (cache_store): Ditto.
+ (cache_remove): Ditto.
+ (forget_host_lookup): No longer necessary, replaced with static
+ function cache_remove.
+
2003-11-14 Hrvoje Niksic <hniksic@xemacs.org>
* main.c: Enable -4 and -6 only if IPv6 is enabled.
diff --git a/src/connect.c b/src/connect.c
index 20538020..d309b272 100644
--- a/src/connect.c
+++ b/src/connect.c
@@ -182,7 +182,7 @@ resolve_bind_address (struct sockaddr *sa)
}
called = 1;
- al = lookup_host_passive (opt.bind_address);
+ al = lookup_host (opt.bind_address, LH_BIND | LH_SILENT);
if (!al)
{
/* #### We should be able to print the error message here. */
@@ -333,10 +333,11 @@ connect_to_host (const char *host, int port)
{
int i, start, end;
struct address_list *al;
+ int lh_flags = 0;
int sock = -1;
again:
- al = lookup_host (host, 0);
+ al = lookup_host (host, lh_flags);
if (!al)
return E_HOST;
@@ -366,7 +367,7 @@ connect_to_host (const char *host, int port)
we were previously able to connect to HOST. That might
indicate that HOST is under dynamic DNS and the addresses
we're connecting to have expired. Resolve it again. */
- forget_host_lookup (host);
+ lh_flags |= LH_REFRESH;
goto again;
}
diff --git a/src/host.c b/src/host.c
index d272c625..6274d0ec 100644
--- a/src/host.c
+++ b/src/host.c
@@ -78,10 +78,6 @@ extern int h_errno;
# endif
#endif
-/* Mapping between known hosts and to lists of their addresses. */
-
-static struct hash_table *host_name_addresses_map;
-
/* Lists of IP addresses that result from running DNS queries. See
lookup_host for details. */
@@ -431,13 +427,40 @@ pretty_print_address (const ip_address *addr)
abort ();
return NULL;
}
+
+/* Simple host cache, used by lookup_host to speed up resolving. The
+ cache doesn't handle TTL because Wget is a fairly short-lived
+ application. Refreshing is attempted when connect fails, though --
+ see connect_to_host. */
-/* Add host name HOST with the address ADDR_TEXT to the cache.
- ADDR_LIST is a NULL-terminated list of addresses, as in struct
- hostent. */
+/* Mapping between known hosts and to lists of their addresses. */
+static struct hash_table *host_name_addresses_map;
+
+
+/* Return the host's resolved addresses from the cache, if
+ available. */
+
+static struct address_list *
+cache_query (const char *host)
+{
+ struct address_list *al;
+ if (!host_name_addresses_map)
+ return NULL;
+ al = hash_table_get (host_name_addresses_map, host);
+ if (al)
+ {
+ DEBUGP (("Found %s in host_name_addresses_map (%p)\n", host, al));
+ ++al->refcount;
+ return al;
+ }
+ return NULL;
+}
+
+/* Cache the DNS lookup of HOST. Subsequent invocations of
+ lookup_host will return the cached value. */
static void
-cache_host_lookup (const char *host, struct address_list *al)
+cache_store (const char *host, struct address_list *al)
{
if (!host_name_addresses_map)
host_name_addresses_map = make_nocase_string_hash_table (0);
@@ -457,20 +480,23 @@ cache_host_lookup (const char *host, struct address_list *al)
#endif
}
-/* Remove HOST from Wget's DNS cache. Does nothing is HOST is not in
+/* Remove HOST from the DNS cache. Does nothing is HOST is not in
the cache. */
-void
-forget_host_lookup (const char *host)
+static void
+cache_remove (const char *host)
{
- struct address_list *al = hash_table_get (host_name_addresses_map, host);
+ struct address_list *al;
+ if (!host_name_addresses_map)
+ return;
+ al = hash_table_get (host_name_addresses_map, host);
if (al)
{
address_list_release (al);
hash_table_remove (host_name_addresses_map, host);
}
}
-
+
/* Look up HOST in DNS and return a list of IP addresses. The
addresses in the list are in the same order in which
gethostbyname/getaddrinfo returned them.
@@ -480,12 +506,20 @@ forget_host_lookup (const char *host)
you want to force lookup, call forget_host_lookup() prior to this
function, or set opt.dns_cache to 0 to globally disable caching.
- If SILENT is non-zero, progress messages are not printed. */
+ FLAGS can be a combination of:
+ LH_SILENT - don't print the "resolving ... done" messages.
+ LH_BIND - resolve addresses for use with bind, which under
+ IPv6 means to use AI_PASSIVE flag to getaddrinfo.
+ Passive lookups are not cached under IPv6.
+ LH_REFRESH - if HOST is cached, remove the entry from the cache
+ and resolve it anew. */
struct address_list *
-lookup_host (const char *host, int silent)
+lookup_host (const char *host, int flags)
{
struct address_list *al = NULL;
+ int silent = flags & LH_SILENT;
+ int use_cache;
#ifndef ENABLE_IPV6
/* If we're not using getaddrinfo, first check if HOST specifies a
@@ -506,20 +540,30 @@ lookup_host (const char *host, int silent)
}
#endif
- /* Try to find the host in the cache. */
+ /* Cache is normally on, but can be turned off with --no-dns-cache.
+ Don't cache passive lookups under IPv6. */
+ use_cache = opt.dns_cache;
+#ifdef ENABLE_IPV6
+ if (flags & LH_BIND)
+ use_cache = 0;
+#endif
- if (host_name_addresses_map)
+ /* Try to find the host in the cache so we don't need to talk to the
+ resolver. If LH_REFRESH is requested, remove HOST from the cache
+ instead. */
+ if (use_cache)
{
- al = hash_table_get (host_name_addresses_map, host);
- if (al)
+ if (!(flags & LH_REFRESH))
{
- DEBUGP (("Found %s in host_name_addresses_map (%p)\n", host, al));
- ++al->refcount;
- return al;
+ al = cache_query (host);
+ if (al)
+ return al;
}
+ else
+ cache_remove (host);
}
- /* No luck with the cache; resolve the host name. */
+ /* No luck with the cache; resolve HOST. */
if (!silent)
logprintf (LOG_VERBOSE, _("Resolving %s... "), host);
@@ -531,9 +575,18 @@ lookup_host (const char *host, int silent)
xzero (hints);
hints.ai_socktype = SOCK_STREAM;
- hints.ai_family = AF_UNSPEC; /* #### should look at opt.ipv4_only
- and opt.ipv6_only */
- hints.ai_flags = 0;
+ hints.ai_family = AF_UNSPEC;
+ if (opt.ipv4_only && !opt.ipv6_only)
+ hints.ai_family = AF_INET;
+ else if (opt.ipv6_only && !opt.ipv4_only)
+ hints.ai_family = AF_INET6;
+
+#ifdef HAVE_GETADDRINFO_AI_ADDRCONFIG
+ /* Use AI_ADDRCONFIG where available. See init.c:defaults(). */
+ hints.ai_flags |= AI_ADDRCONFIG;
+#endif
+ if (flags & LH_BIND)
+ hints.ai_flags |= AI_PASSIVE;
err = getaddrinfo_with_timeout (host, NULL, &hints, &res, opt.dns_timeout);
if (err != 0 || res == NULL)
@@ -547,7 +600,8 @@ lookup_host (const char *host, int silent)
freeaddrinfo (res);
if (!al)
{
- logprintf (LOG_VERBOSE, _("failed: No IPv4/IPv6 addresses.\n"));
+ logprintf (LOG_VERBOSE,
+ _("failed: No IPv4/IPv6 addresses for host.\n"));
return NULL;
}
}
@@ -590,49 +644,11 @@ lookup_host (const char *host, int silent)
}
/* Cache the lookup information. */
- if (opt.dns_cache)
- cache_host_lookup (host, al);
+ if (use_cache)
+ cache_store (host, al);
return al;
}
-
-/* Resolve HOST to get an address for use with bind(2). Do *not* use
- this for sockets to be used with connect(2).
-
- This is a function separate from lookup_host because the results it
- returns are different -- it uses the AI_PASSIVE flag to
- getaddrinfo. Because of this distinction, it doesn't store the
- results in the cache. It prints nothing and implements no timeouts
- because it should normally only be used with local addresses
- (typically "localhost" or numeric addresses of different local
- interfaces.)
-
- Without IPv6, this function just calls lookup_host. */
-
-struct address_list *
-lookup_host_passive (const char *host)
-{
-#ifdef ENABLE_IPV6
- struct address_list *al = NULL;
- int err;
- struct addrinfo hints, *res;
-
- xzero (hints);
- hints.ai_socktype = SOCK_STREAM;
- hints.ai_family = AF_UNSPEC; /* #### should look at opt.ipv4_only
- and opt.ipv6_only */
- hints.ai_flags = AI_PASSIVE;
-
- err = getaddrinfo (host, NULL, &hints, &res);
- if (err != 0 || res == NULL)
- return NULL;
- al = address_list_from_addrinfo (res);
- freeaddrinfo (res);
- return al;
-#else
- return lookup_host (host, 1);
-#endif
-}
/* Determine whether a URL is acceptable to be followed, according to
a list of domains to accept. */
diff --git a/src/host.h b/src/host.h
index 7790c8b8..4f37d2fe 100644
--- a/src/host.h
+++ b/src/host.h
@@ -90,11 +90,12 @@ typedef struct {
#define ADDRESS_IPV6_DATA(x) ((void *)&(x)->u.ipv6.addr)
#define ADDRESS_IPV6_SCOPE(x) ((x)->u.ipv6.scope_id)
-/* Function declarations */
+enum {
+ LH_SILENT = 1,
+ LH_BIND = 2,
+ LH_REFRESH = 4
+};
struct address_list *lookup_host PARAMS ((const char *, int));
-struct address_list *lookup_host_passive PARAMS ((const char *));
-
-void forget_host_lookup PARAMS ((const char *));
void address_list_get_bounds PARAMS ((const struct address_list *,
int *, int *));
diff --git a/src/init.c b/src/init.c
index 74e49cc1..e93d6226 100644
--- a/src/init.c
+++ b/src/init.c
@@ -166,11 +166,11 @@ static struct {
{ "ignorelength", &opt.ignore_length, cmd_boolean },
{ "ignoretags", &opt.ignore_tags, cmd_vector },
{ "includedirectories", &opt.includes, cmd_directory_vector },
- { "input", &opt.input_filename, cmd_file },
#ifdef ENABLE_IPV6
{ "inet4only", &opt.ipv4_only, cmd_boolean },
{ "inet6only", &opt.ipv6_only, cmd_boolean },
#endif
+ { "input", &opt.input_filename, cmd_file },
{ "keepsessioncookies", &opt.keep_session_cookies, cmd_boolean },
{ "killlonger", &opt.kill_longer, cmd_boolean },
{ "limitrate", &opt.limit_rate, cmd_bytes },