/* Handling of recursive HTTP retrieving.
- Copyright (C) 1995, 1996, 1997 Free Software Foundation, Inc.
+ Copyright (C) 1995, 1996, 1997, 2000 Free Software Foundation, Inc.
This file is part of Wget.
#include "ftp.h"
#include "fnmatch.h"
#include "host.h"
+#include "hash.h"
extern char *version_string;
#define ROBOTS_FILENAME "robots.txt"
-/* #### Many of these lists should really be hashtables! */
-
-/* List of downloaded URLs. */
-static urlpos *urls_downloaded;
+static struct hash_table *dl_file_url_map;
+static struct hash_table *dl_url_file_map;
/* List of HTML URLs. */
static slist *urls_html;
/* List of undesirable-to-load URLs. */
-static slist *ulist;
+static struct hash_table *undesirable_urls;
/* List of forbidden locations. */
static char **forbidden = NULL;
void
recursive_cleanup (void)
{
- free_slist (ulist);
- ulist = NULL;
+ if (undesirable_urls)
+ {
+ string_set_free (undesirable_urls);
+ undesirable_urls = NULL;
+ }
+ if (dl_file_url_map)
+ {
+ free_keys_and_values (dl_file_url_map);
+ hash_table_destroy (dl_file_url_map);
+ dl_file_url_map = NULL;
+ }
+ if (dl_url_file_map)
+ {
+ free_keys_and_values (dl_url_file_map);
+ hash_table_destroy (dl_url_file_map);
+ dl_url_file_map = NULL;
+ }
+ undesirable_urls = NULL;
free_vec (forbidden);
forbidden = NULL;
- free_slist (urls_html);
+ slist_free (urls_html);
urls_html = NULL;
- free_urlpos (urls_downloaded);
- urls_downloaded = NULL;
FREE_MAYBE (base_dir);
FREE_MAYBE (robots_host);
first_time = 1;
char *constr, *filename, *newloc;
char *canon_this_url = NULL;
int dt, inl, dash_p_leaf_HTML = FALSE;
+ int meta_disallow_follow;
int this_url_ftp; /* See below the explanation */
uerr_t err;
struct urlinfo *rurl;
/* Cache the current URL in the list. */
if (first_time)
{
- ulist = add_slist (ulist, this_url, 0);
- urls_downloaded = NULL;
+ /* These three operations need to be done only once per Wget
+ run. They should probably be at a different location. */
+ if (!undesirable_urls)
+ undesirable_urls = make_string_hash_table (0);
+ if (!dl_file_url_map)
+ dl_file_url_map = make_string_hash_table (0);
+ if (!dl_url_file_map)
+ dl_url_file_map = make_string_hash_table (0);
+
+ hash_table_clear (undesirable_urls);
+ string_set_add (undesirable_urls, this_url);
+ hash_table_clear (dl_file_url_map);
+ hash_table_clear (dl_url_file_map);
urls_html = NULL;
- /* Enter this_url to the slist, in original and "enhanced" form. */
+ /* Enter this_url to the hash table, in original and "enhanced" form. */
u = newurl ();
err = parseurl (this_url, u, 0);
if (err == URLOK)
{
- ulist = add_slist (ulist, u->url, 0);
- urls_downloaded = add_url (urls_downloaded, u->url, file);
- urls_html = add_slist (urls_html, file, NOSORT);
+ string_set_add (undesirable_urls, u->url);
+ hash_table_put (dl_file_url_map, xstrdup (file), xstrdup (u->url));
+ hash_table_put (dl_url_file_map, xstrdup (u->url), xstrdup (file));
+ urls_html = slist_prepend (urls_html, file);
if (opt.no_parent)
base_dir = xstrdup (u->dir); /* Set the base dir. */
/* Set the canonical this_url to be sent as referer. This
/* Get the URL-s from an HTML file: */
url_list = get_urls_html (file, canon_this_url ? canon_this_url : this_url,
- 0, dash_p_leaf_HTML);
+ dash_p_leaf_HTML, &meta_disallow_follow);
+
+ if (opt.use_robots && meta_disallow_follow)
+ {
+ /* The META tag says we are not to follow this file. Respect
+ that. */
+ free_urlpos (url_list);
+ url_list = NULL;
+ }
/* Decide what to do with each of the URLs. A URL will be loaded if
it meets several requirements, discussed later. */
the list. */
/* inl is set if the URL we are working on (constr) is stored in
- ulist. Using it is crucial to avoid the incessant calls to
- in_slist, which is quite slow. */
- inl = in_slist (ulist, constr);
+ undesirable_urls. Using it is crucial to avoid unnecessary
+ repeated continuous hits to the hash table. */
+ inl = string_set_exists (undesirable_urls, constr);
/* If it is FTP, and FTP is not followed, chuck it out. */
if (!inl)
if (u->proto == URLFTP && !opt.follow_ftp && !this_url_ftp)
{
DEBUGP (("Uh, it is FTP but i'm not in the mood to follow FTP.\n"));
- ulist = add_slist (ulist, constr, 0);
+ string_set_add (undesirable_urls, constr);
inl = 1;
}
/* If it is absolute link and they are not followed, chuck it
out. */
if (!inl && u->proto != URLFTP)
- if (opt.relative_only && !(cur_url->flags & URELATIVE))
+ if (opt.relative_only && !cur_url->link_relative_p)
{
DEBUGP (("It doesn't really look like a relative link.\n"));
- ulist = add_slist (ulist, constr, 0);
+ string_set_add (undesirable_urls, constr);
inl = 1;
}
/* If its domain is not to be accepted/looked-up, chuck it out. */
if (!accept_domain (u))
{
DEBUGP (("I don't like the smell of that domain.\n"));
- ulist = add_slist (ulist, constr, 0);
+ string_set_add (undesirable_urls, constr);
inl = 1;
}
/* Check for parent directory. */
{
/* Failing that too, kill the URL. */
DEBUGP (("Trying to escape parental guidance with no_parent on.\n"));
- ulist = add_slist (ulist, constr, 0);
+ string_set_add (undesirable_urls, constr);
inl = 1;
}
freeurl (ut, 1);
if (!accdir (u->dir, ALLABS))
{
DEBUGP (("%s (%s) is excluded/not-included.\n", constr, u->dir));
- ulist = add_slist (ulist, constr, 0);
+ string_set_add (undesirable_urls, constr);
inl = 1;
}
}
{
DEBUGP (("%s (%s) does not match acc/rej rules.\n",
constr, u->file));
- ulist = add_slist (ulist, constr, 0);
+ string_set_add (undesirable_urls, constr);
inl = 1;
}
}
}
free (constr);
constr = xstrdup (u->url);
- inl = in_slist (ulist, constr);
+ string_set_add (undesirable_urls, constr);
if (!inl && !((u->proto == URLFTP) && !this_url_ftp))
if (!opt.spanhost && this_url && !same_host (this_url, constr))
{
DEBUGP (("This is not the same hostname as the parent's.\n"));
- ulist = add_slist (ulist, constr, 0);
+ string_set_add (undesirable_urls, constr);
inl = 1;
}
}
{
DEBUGP (("Stuffing %s because %s forbids it.\n", this_url,
ROBOTS_FILENAME));
- ulist = add_slist (ulist, constr, 0);
+ string_set_add (undesirable_urls, constr);
inl = 1;
}
}
{
DEBUGP (("I've decided to load it -> "));
/* Add it to the list of already-loaded URL-s. */
- ulist = add_slist (ulist, constr, 0);
+ string_set_add (undesirable_urls, constr);
/* Automatically followed FTPs will *not* be downloaded
recursively. */
if (u->proto == URLFTP)
{
if (dt & RETROKF)
{
- urls_downloaded = add_url (urls_downloaded, constr, filename);
+ hash_table_put (dl_file_url_map,
+ xstrdup (filename), xstrdup (constr));
+ hash_table_put (dl_url_file_map,
+ xstrdup (constr), xstrdup (filename));
/* If the URL is HTML, note it. */
if (dt & TEXTHTML)
- urls_html = add_slist (urls_html, filename, NOSORT);
+ urls_html = slist_prepend (urls_html, filename);
}
}
/* If there was no error, and the type is text/html, parse
store the local filename. */
if (opt.convert_links && (dt & RETROKF) && (filename != NULL))
{
- cur_url->flags |= UABS2REL;
+ cur_url->convert = CO_CONVERT_TO_RELATIVE;
cur_url->local_name = xstrdup (filename);
}
}
- DEBUGP (("%s already in list, so we don't load.\n", constr));
+ else
+ DEBUGP (("%s already in list, so we don't load.\n", constr));
/* Free filename and constr. */
FREE_MAYBE (filename);
FREE_MAYBE (constr);
/* Increment the pbuf for the appropriate size. */
}
if (opt.convert_links && !opt.delete_after)
+ /* This is merely the first pass: the links that have been
+ successfully downloaded are converted. In the second pass,
+ convert_all_links() will also convert those links that have NOT
+ been downloaded to their canonical form. */
convert_links (file, url_list);
/* Free the linked list of URL-s. */
free_urlpos (url_list);
return RETROK;
}
\f
-/* Simple calls to convert_links will often fail because only the
- downloaded files are converted, and Wget cannot know which files
- will be converted in the future. So, if we have file fileone.html
- with:
+/* convert_links() is called from recursive_retrieve() after we're
+ done with an HTML file. This call to convert_links is not complete
+ because it converts only the downloaded files, and Wget cannot know
+ which files will be downloaded afterwards. So, if we have file
+ fileone.html with:
- <a href=/c/something.gif>
+ <a href="/c/something.gif">
and /c/something.gif was not downloaded because it exceeded the
recursion depth, the reference will *not* be changed.
void
convert_all_links (void)
{
- uerr_t res;
- urlpos *l1, *l2, *urls;
- struct urlinfo *u;
slist *html;
- urlpos *urlhtml;
+
+ /* Destructively reverse urls_html to get it in the right order.
+ recursive_retrieve() used slist_prepend() consistently. */
+ urls_html = slist_nreverse (urls_html);
for (html = urls_html; html; html = html->next)
{
+ urlpos *urls, *cur_url;
+ char *url;
+
DEBUGP (("Rescanning %s\n", html->string));
/* Determine the URL of the HTML file. get_urls_html will need
it. */
- for (urlhtml = urls_downloaded; urlhtml; urlhtml = urlhtml->next)
- if (!strcmp (urlhtml->local_name, html->string))
- break;
- if (urlhtml)
- DEBUGP (("It should correspond to %s.\n", urlhtml->url));
+ url = hash_table_get (dl_file_url_map, html->string);
+ if (url)
+ DEBUGP (("It should correspond to %s.\n", url));
else
DEBUGP (("I cannot find the corresponding URL.\n"));
/* Parse the HTML file... */
- urls = get_urls_html (html->string, urlhtml ? urlhtml->url : NULL, 1,
- FALSE);
- if (!urls)
- continue;
- for (l1 = urls; l1; l1 = l1->next)
+ urls = get_urls_html (html->string, url, FALSE, NULL);
+ /* We don't respect meta_disallow_follow here because, even if
+ the file is not followed, we might still want to convert the
+ links that have been followed from other files. */
+ for (cur_url = urls; cur_url; cur_url = cur_url->next)
{
+ char *local_name;
+
/* The URL must be in canonical form to be compared. */
- u = newurl ();
- res = parseurl (l1->url, u, 0);
+ struct urlinfo *u = newurl ();
+ uerr_t res = parseurl (cur_url->url, u, 0);
if (res != URLOK)
{
freeurl (u, 1);
}
/* We decide the direction of conversion according to whether
a URL was downloaded. Downloaded URLs will be converted
- ABS2REL, whereas non-downloaded will be converted REL2ABS.
- Note: not yet implemented; only ABS2REL works. */
- for (l2 = urls_downloaded; l2; l2 = l2->next)
- if (!strcmp (l2->url, u->url))
- {
- DEBUGP (("%s flagged for conversion, local %s\n",
- l2->url, l2->local_name));
- break;
- }
- /* Clear the flags. */
- l1->flags &= ~ (UABS2REL | UREL2ABS);
+ ABS2REL, whereas non-downloaded will be converted REL2ABS. */
+ local_name = hash_table_get (dl_url_file_map, u->url);
+ if (local_name)
+ DEBUGP (("%s marked for conversion, local %s\n",
+ u->url, local_name));
/* Decide on the conversion direction. */
- if (l2)
+ if (local_name)
{
- l1->flags |= UABS2REL;
- l1->local_name = xstrdup (l2->local_name);
+ /* We've downloaded this URL. Convert it to relative
+ form. We do this even if the URL already is in
+ relative form, because our directory structure may
+ not be identical to that on the server (think `-nd',
+ `--cut-dirs', etc.) */
+ cur_url->convert = CO_CONVERT_TO_RELATIVE;
+ cur_url->local_name = xstrdup (local_name);
}
else
{
- l1->flags |= UREL2ABS;
- l1->local_name = NULL;
+ /* We haven't downloaded this URL. If it's not already
+ complete (including a full host name), convert it to
+ that form, so it can be reached while browsing this
+ HTML locally. */
+ if (!cur_url->link_complete_p)
+ cur_url->convert = CO_CONVERT_TO_COMPLETE;
+ cur_url->local_name = NULL;
}
freeurl (u, 1);
}
while ((line = read_whole_line (fp)))
{
len = strlen (line);
- /* Destroy <CR> if there is one. */
+ /* Destroy <CR><LF> if present. */
+ if (len && line[len - 1] == '\n')
+ line[--len] = '\0';
if (len && line[len - 1] == '\r')
- line[len - 1] = '\0';
+ line[--len] = '\0';
/* According to specifications, optional space may be at the
end... */
DEBUGP (("Line: %s\n", line));