/* HTTP support.
- Copyright (C) 1996-2006 Free Software Foundation, Inc.
+ Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
+ 2004, 2005, 2006, 2007 Free Software Foundation, Inc.
This file is part of GNU Wget.
only hurts us. */
request_remove_header (req, "Authorization");
}
- }
-
- if (sock < 0)
- {
- /* In its current implementation, persistent_available_p will
- look up conn->host in some cases. If that lookup failed, we
- don't need to bother with connect_to_host. */
- if (host_lookup_failed)
+ else if (host_lookup_failed)
{
request_free (req);
+ logprintf(LOG_NOTQUIET,
+ _("%s: unable to resolve host address `%s'\n"),
+ exec_name, relevant->host);
return HOSTERR;
}
+ }
+ if (sock < 0)
+ {
sock = connect_to_host (conn->host, conn->port);
if (sock == E_HOST)
{
if (has_html_suffix_p (hs->local_file))
*dt |= TEXTHTML;
- return RETROK;
+ return RETRUNNEEDED;
}
else if (!ALLOW_CLOBBER)
{
errno = 0;
parsed = str_to_wgint (hdrval, NULL, 10);
if (parsed == WGINT_MAX && errno == ERANGE)
- /* Out of range.
- #### If Content-Length is out of range, it most likely
- means that the file is larger than 2G and that we're
- compiled without LFS. In that case we should probably
- refuse to even attempt to download the file. */
- contlen = -1;
+ {
+ /* Out of range.
+ #### If Content-Length is out of range, it most likely
+ means that the file is larger than 2G and that we're
+ compiled without LFS. In that case we should probably
+ refuse to even attempt to download the file. */
+ contlen = -1;
+ }
+ else if (parsed < 0)
+ {
+ /* Negative Content-Length; nonsensical, so we can't
+ assume any information about the content to receive. */
+ contlen = -1;
+ }
else
contlen = parsed;
}
{
int count;
bool got_head = false; /* used for time-stamping and filename detection */
+ bool time_came_from_head = false;
bool got_name = false;
char *tms;
const char *tmrate;
uerr_t err, ret = TRYLIMEXC;
time_t tmr = -1; /* remote time-stamp */
- wgint local_size = 0; /* the size of the local file */
struct http_stat hstat; /* HTTP status */
struct_stat st;
+ bool send_head_first = true;
/* Assert that no value for *LOCAL_FILE was passed. */
assert (local_file == NULL || *local_file == NULL);
/* Reset the document type. */
*dt = 0;
+ /* Skip preliminary HEAD request if we're not in spider mode AND
+ * if -O was given or HTTP Content-Disposition support is disabled. */
+ if (!opt.spider
+ && (got_name || !opt.content_disposition))
+ send_head_first = false;
+
+ /* Send preliminary HEAD request if -N is given and we have an existing
+ * destination file. */
+ if (opt.timestamping
+ && !opt.content_disposition
+ && file_exists_p (url_file_name (u)))
+ send_head_first = true;
+
/* THE loop */
do
{
sleep_between_retrievals (count);
/* Get the current time string. */
- tms = time_str (time (NULL));
+ tms = datetime_str (time (NULL));
if (opt.spider && !got_head)
logprintf (LOG_VERBOSE, _("\
/* Print fetch message, if opt.verbose. */
if (opt.verbose)
{
- char *hurl = url_string (u, true);
+ char *hurl = url_string (u, URL_AUTH_HIDE_PASSWD);
if (count > 1)
{
/* Default document type is empty. However, if spider mode is
on or time-stamping is employed, HEAD_ONLY commands is
encoded within *dt. */
- if (((opt.spider || opt.timestamping) && !got_head)
- || (opt.always_rest && !got_name))
+ if (send_head_first && !got_head)
*dt |= HEAD_ONLY;
else
*dt &= ~HEAD_ONLY;
err = gethttp (u, &hstat, dt, proxy);
/* Time? */
- tms = time_str (time (NULL));
+ tms = datetime_str (time (NULL));
/* Get the new location (with or without the redirection). */
if (hstat.newloc)
*newloc = xstrdup (hstat.newloc);
-
+
switch (err)
{
case HERR: case HEOF: case CONSOCKERR: case CONCLOSED:
/* All possibilities should have been exhausted. */
abort ();
}
-
+
if (!(*dt & RETROKF))
{
char *hurl = NULL;
if (!opt.verbose)
{
/* #### Ugly ugly ugly! */
- hurl = url_string (u, true);
+ hurl = url_string (u, URL_AUTH_HIDE_PASSWD);
logprintf (LOG_NONVERBOSE, "%s:\n", hurl);
}
+
+ /* Fall back to GET if HEAD fails with a 500 or 501 error code. */
+ if (*dt & HEAD_ONLY
+ && (hstat.statcode == 500 || hstat.statcode == 501))
+ {
+ got_head = true;
+ continue;
+ }
/* Maybe we should always keep track of broken links, not just in
* spider mode. */
- if (opt.spider)
+ else if (opt.spider)
{
/* #### Again: ugly ugly ugly! */
if (!hurl)
- hurl = url_string (u, true);
+ hurl = url_string (u, URL_AUTH_HIDE_PASSWD);
nonexisting_url (hurl);
logprintf (LOG_NOTQUIET, _("\
Remote file does not exist -- broken link!!!\n"));
}
/* Did we get the time-stamp? */
- if (!got_head)
+ if (send_head_first && !got_head)
{
bool restart_loop = false;
if (tmr == (time_t) (-1))
logputs (LOG_VERBOSE, _("\
Last-modified header invalid -- time-stamp ignored.\n"));
+ if (*dt & HEAD_ONLY)
+ time_came_from_head = true;
}
/* The time-stamping section. */
{
logprintf (LOG_VERBOSE, _("\
The sizes do not match (local %s) -- retrieving.\n"),
- number_to_static_string (local_size));
+ number_to_static_string (hstat.orig_file_size));
}
}
else
restart_loop = true;
}
- if (opt.always_rest)
- {
- got_name = true;
- restart_loop = true;
- }
-
if (opt.spider)
{
if (opt.recursive)
{
logprintf (LOG_VERBOSE, _("\
Remote file exists but does not contain any link -- not retrieving.\n\n"));
- ret = RETRUNNEEDED;
+ ret = RETROK; /* RETRUNNEEDED is not for caller. */
goto exit;
}
}
{
logprintf (LOG_VERBOSE, _("\
Remote file exists but recursion is disabled -- not retrieving.\n\n"));
- ret = RETRUNNEEDED;
+ ret = RETROK; /* RETRUNNEEDED is not for caller. */
goto exit;
}
}
+ if (send_head_first)
+ {
+ got_name = true;
+ restart_loop = true;
+ }
+
got_head = true; /* no more time-stamping */
*dt &= ~HEAD_ONLY;
count = 0; /* the retrieve count for HEAD is reset */
else
fl = hstat.local_file;
if (fl)
- touch (fl, tmr);
+ {
+ time_t newtmr = -1;
+ /* Reparse time header, in case it's changed. */
+ if (time_came_from_head
+ && hstat.remote_time && hstat.remote_time[0])
+ {
+ newtmr = http_atotm (hstat.remote_time);
+ if (newtmr != -1)
+ tmr = newtmr;
+ }
+ touch (fl, tmr);
+ }
}
/* End of time-stamping section. */