#include "fnmatch.h"
#include "host.h"
#include "hash.h"
+#include "res.h"
#ifndef errno
extern int errno;
extern char *version_string;
-#define ROBOTS_FILENAME "robots.txt"
-
static struct hash_table *dl_file_url_map;
static struct hash_table *dl_url_file_map;
/* List of undesirable-to-load URLs. */
static struct hash_table *undesirable_urls;
-/* List of forbidden locations. */
-static char **forbidden = NULL;
-
/* Current recursion depth. */
static int depth;
/* Base directory we're recursing from (used by no_parent). */
static char *base_dir;
-/* The host name for which we last checked robots. */
-static char *robots_host;
-
static int first_time = 1;
-/* Construct the robots URL. */
-static struct urlinfo *robots_url PARAMS ((const char *, const char *));
-static uerr_t retrieve_robots PARAMS ((const char *, const char *));
-static char **parse_robots PARAMS ((const char *));
-static int robots_match PARAMS ((struct urlinfo *, char **));
-
/* Cleanup the data structures associated with recursive retrieving
(the variables above). */
dl_url_file_map = NULL;
}
undesirable_urls = NULL;
- free_vec (forbidden);
- forbidden = NULL;
slist_free (downloaded_html_files);
downloaded_html_files = NULL;
FREE_MAYBE (base_dir);
- FREE_MAYBE (robots_host);
first_time = 1;
}
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;
urlpos *url_list, *cur_url;
- char *rfile; /* For robots */
- struct urlinfo *u;
+ struct url *u;
assert (this_url != NULL);
assert (file != NULL);
hash_table_clear (undesirable_urls);
string_set_add (undesirable_urls, this_url);
/* Enter this_url to the hash table, in original and "enhanced" form. */
- u = newurl ();
- err = parseurl (this_url, u, 0);
- if (err == URLOK)
+ u = url_parse (this_url, NULL);
+ if (u)
{
string_set_add (undesirable_urls, u->url);
if (opt.no_parent)
DEBUGP (("Double yuck! The *base* URL is broken.\n"));
base_dir = NULL;
}
- freeurl (u, 1);
+ url_free (u);
depth = 1;
- robots_host = NULL;
- forbidden = NULL;
first_time = 0;
}
else
that the retrieval is done through proxy. In that case, FTP
links will be followed by default and recursion will not be
turned off when following them. */
- this_url_ftp = (urlproto (this_url) == URLFTP);
+ this_url_ftp = (url_scheme (this_url) == SCHEME_FTP);
/* Get the URL-s from an HTML file: */
url_list = get_urls_html (file, canon_this_url ? canon_this_url : this_url,
break;
/* Parse the URL for convenient use in other functions, as well
as to get the optimized form. It also checks URL integrity. */
- u = newurl ();
- if (parseurl (cur_url->url, u, 0) != URLOK)
+ u = url_parse (cur_url->url, NULL);
+ if (!u)
{
DEBUGP (("Yuck! A bad URL.\n"));
- freeurl (u, 1);
- continue;
- }
- if (u->proto == URLFILE)
- {
- DEBUGP (("Nothing to do with file:// around here.\n"));
- freeurl (u, 1);
continue;
}
assert (u->url != NULL);
/* If it is FTP, and FTP is not followed, chuck it out. */
if (!inl)
- if (u->proto == URLFTP && !opt.follow_ftp && !this_url_ftp)
+ if (u->scheme == SCHEME_FTP && !opt.follow_ftp && !this_url_ftp)
{
DEBUGP (("Uh, it is FTP but i'm not in the mood to follow FTP.\n"));
string_set_add (undesirable_urls, constr);
}
/* If it is absolute link and they are not followed, chuck it
out. */
- if (!inl && u->proto != URLFTP)
+ if (!inl && u->scheme != SCHEME_FTP)
if (opt.relative_only && !cur_url->link_relative_p)
{
DEBUGP (("It doesn't really look like a relative link.\n"));
if (!inl && opt.no_parent
/* If the new URL is FTP and the old was not, ignore
opt.no_parent. */
- && !(!this_url_ftp && u->proto == URLFTP))
+ && !(!this_url_ftp && u->scheme == SCHEME_FTP))
{
/* Check for base_dir first. */
if (!(base_dir && frontcmp (base_dir, u->dir)))
{
/* Failing that, check for parent dir. */
- struct urlinfo *ut = newurl ();
- if (parseurl (this_url, ut, 0) != URLOK)
+ struct url *ut = url_parse (this_url, NULL);
+ if (!ut)
DEBUGP (("Double yuck! The *base* URL is broken.\n"));
else if (!frontcmp (ut->dir, u->dir))
{
string_set_add (undesirable_urls, constr);
inl = 1;
}
- freeurl (ut, 1);
+ url_free (ut);
}
}
/* If the file does not match the acceptance list, or is on the
if (!inl)
{
if (!opt.simple_check)
- opt_url (u);
+ {
+ /* Find the "true" host. */
+ char *host = realhost (u->host);
+ xfree (u->host);
+ u->host = host;
+
+ /* Refresh the printed representation of the URL. */
+ xfree (u->url);
+ u->url = url_string (u, 0);
+ }
else
{
char *p;
for (p = u->host; *p; p++)
*p = TOLOWER (*p);
xfree (u->url);
- u->url = str_url (u, 0);
+ u->url = url_string (u, 0);
}
xfree (constr);
constr = xstrdup (u->url);
/* This line is bogus. */
/*string_set_add (undesirable_urls, constr);*/
- if (!inl && !((u->proto == URLFTP) && !this_url_ftp))
+ if (!inl && !((u->scheme == SCHEME_FTP) && !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"));
}
}
/* What about robots.txt? */
- if (!inl && opt.use_robots && u->proto == URLHTTP)
+ if (!inl && opt.use_robots && u->scheme == SCHEME_FTP)
{
- /* Since Wget knows about only one set of robot rules at a
- time, /robots.txt must be reloaded whenever a new host is
- accessed.
-
- robots_host holds the host the current `forbid' variable
- is assigned to. */
- if (!robots_host || !same_host (robots_host, u->host))
+ struct robot_specs *specs = res_get_specs (u->host, u->port);
+ if (!specs)
{
- FREE_MAYBE (robots_host);
- /* Now make robots_host the new host, no matter what the
- result will be. So if there is no /robots.txt on the
- site, Wget will not retry getting robots all the
- time. */
- robots_host = xstrdup (u->host);
- free_vec (forbidden);
- forbidden = NULL;
- err = retrieve_robots (constr, ROBOTS_FILENAME);
- if (err == ROBOTSOK)
+ char *rfile;
+ if (res_retrieve_file (constr, &rfile))
{
- rurl = robots_url (constr, ROBOTS_FILENAME);
- rfile = url_filename (rurl);
- forbidden = parse_robots (rfile);
- freeurl (rurl, 1);
+ specs = res_parse_from_file (rfile);
xfree (rfile);
}
+ else
+ {
+ /* If we cannot get real specs, at least produce
+ dummy ones so that we can register them and stop
+ trying to retrieve them. */
+ specs = res_parse ("", 0);
+ }
+ res_register_specs (u->host, u->port, specs);
}
- /* Now that we have (or don't have) robots, we can check for
- them. */
- if (!robots_match (u, forbidden))
+ /* Now that we have (or don't have) robots.txt specs, we can
+ check what they say. */
+ if (!res_match_path (specs, u->path))
{
- DEBUGP (("Stuffing %s because %s forbids it.\n", this_url,
- ROBOTS_FILENAME));
+ DEBUGP (("Not following %s because robots.txt forbids it.\n",
+ constr));
string_set_add (undesirable_urls, constr);
inl = 1;
}
string_set_add (undesirable_urls, constr);
/* Automatically followed FTPs will *not* be downloaded
recursively. */
- if (u->proto == URLFTP)
+ if (u->scheme == SCHEME_FTP)
{
/* Don't you adore side-effects? */
opt.recursive = 0;
/* Retrieve it. */
retrieve_url (constr, &filename, &newloc,
canon_this_url ? canon_this_url : this_url, &dt);
- if (u->proto == URLFTP)
+ if (u->scheme == SCHEME_FTP)
{
/* Restore... */
opt.recursive = 1;
/* Free filename and constr. */
FREE_MAYBE (filename);
FREE_MAYBE (constr);
- freeurl (u, 1);
+ url_free (u);
/* Increment the pbuf for the appropriate size. */
}
if (opt.convert_links && !opt.delete_after)
char *local_name;
/* The URL must be in canonical form to be compared. */
- struct urlinfo *u = newurl ();
- uerr_t res = parseurl (cur_url->url, u, 0);
- if (res != URLOK)
- {
- freeurl (u, 1);
- continue;
- }
+ struct url *u = url_parse (cur_url->url, NULL);
+ if (!u)
+ continue;
/* 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. */
cur_url->convert = CO_CONVERT_TO_COMPLETE;
cur_url->local_name = NULL;
}
- freeurl (u, 1);
+ url_free (u);
}
/* Convert the links in the file. */
convert_links (html->string, urls);
free_urlpos (urls);
}
}
-\f
-/* Robots support. */
-
-/* Construct the robots URL. */
-static struct urlinfo *
-robots_url (const char *url, const char *robots_filename)
-{
- struct urlinfo *u = newurl ();
- uerr_t err;
-
- err = parseurl (url, u, 0);
- assert (err == URLOK && u->proto == URLHTTP);
- xfree (u->file);
- xfree (u->dir);
- xfree (u->url);
- u->dir = xstrdup ("");
- u->file = xstrdup (robots_filename);
- u->url = str_url (u, 0);
- return u;
-}
-
-/* Retrieves the robots_filename from the root server directory, if
- possible. Returns ROBOTSOK if robots were retrieved OK, and
- NOROBOTS if robots could not be retrieved for any reason. */
-static uerr_t
-retrieve_robots (const char *url, const char *robots_filename)
-{
- int dt;
- uerr_t err;
- struct urlinfo *u;
-
- u = robots_url (url, robots_filename);
- logputs (LOG_VERBOSE, _("Loading robots.txt; please ignore errors.\n"));
- err = retrieve_url (u->url, NULL, NULL, NULL, &dt);
- freeurl (u, 1);
- if (err == RETROK)
- return ROBOTSOK;
- else
- return NOROBOTS;
-}
-
-/* Parse the robots_filename and return the disallowed path components
- in a malloc-ed vector of character pointers.
-
- It should be fully compliant with the syntax as described in the
- file norobots.txt, adopted by the robots mailing list
- (robots@webcrawler.com). */
-static char **
-parse_robots (const char *robots_filename)
-{
- FILE *fp;
- char **entries;
- char *line, *cmd, *str, *p;
- char *base_version, *version;
- int num, i;
- int wget_matched; /* is the part meant for Wget? */
-
- entries = NULL;
-
- num = 0;
- fp = fopen (robots_filename, "rb");
- if (!fp)
- return NULL;
-
- /* Kill version number. */
- if (opt.useragent)
- {
- STRDUP_ALLOCA (base_version, opt.useragent);
- STRDUP_ALLOCA (version, opt.useragent);
- }
- else
- {
- int len = 10 + strlen (version_string);
- base_version = (char *)alloca (len);
- sprintf (base_version, "Wget/%s", version_string);
- version = (char *)alloca (len);
- sprintf (version, "Wget/%s", version_string);
- }
- for (p = version; *p; p++)
- *p = TOLOWER (*p);
- for (p = base_version; *p && *p != '/'; p++)
- *p = TOLOWER (*p);
- *p = '\0';
-
- /* Setting this to 1 means that Wget considers itself under
- restrictions by default, even if the User-Agent field is not
- present. However, if it finds the user-agent set to anything
- other than Wget, the rest will be ignored (up to the following
- User-Agent field). Thus you may have something like:
-
- Disallow: 1
- Disallow: 2
- User-Agent: stupid-robot
- Disallow: 3
- Disallow: 4
- User-Agent: Wget*
- Disallow: 5
- Disallow: 6
- User-Agent: *
- Disallow: 7
-
- In this case the 1, 2, 5, 6 and 7 disallow lines will be
- stored. */
- wget_matched = 1;
- while ((line = read_whole_line (fp)))
- {
- int len = strlen (line);
- /* Destroy <CR><LF> if present. */
- if (len && line[len - 1] == '\n')
- line[--len] = '\0';
- if (len && line[len - 1] == '\r')
- line[--len] = '\0';
- /* According to specifications, optional space may be at the
- end... */
- DEBUGP (("Line: %s\n", line));
- /* Skip spaces. */
- for (cmd = line; *cmd && ISSPACE (*cmd); cmd++);
- if (!*cmd)
- {
- xfree (line);
- DEBUGP (("(chucked out)\n"));
- continue;
- }
- /* Look for ':'. */
- for (str = cmd; *str && *str != ':'; str++);
- if (!*str)
- {
- xfree (line);
- DEBUGP (("(chucked out)\n"));
- continue;
- }
- /* Zero-terminate the command. */
- *str++ = '\0';
- /* Look for the string beginning... */
- for (; *str && ISSPACE (*str); str++);
- /* Look for comments or trailing spaces and kill them off. */
- for (p = str; *p; p++)
- if (*p && ISSPACE (*p) && ((*(p + 1) == '#') || (*(p + 1) == '\0')))
- {
- /* We have found either a shell-style comment `<sp>+#' or some
- trailing spaces. Now rewind to the beginning of the spaces
- and place '\0' there. */
- while (p > str && ISSPACE (*p))
- --p;
- if (p == str)
- *p = '\0';
- else
- *(p + 1) = '\0';
- break;
- }
- if (!strcasecmp (cmd, "User-agent"))
- {
- int match = 0;
- /* Lowercase the agent string. */
- for (p = str; *p; p++)
- *p = TOLOWER (*p);
- /* If the string is `*', it matches. */
- if (*str == '*' && !*(str + 1))
- match = 1;
- else
- {
- /* If the string contains wildcards, we'll run it through
- fnmatch(). */
- if (has_wildcards_p (str))
- {
- /* If the string contains '/', compare with the full
- version. Else, compare it to base_version. */
- if (strchr (str, '/'))
- match = !fnmatch (str, version, 0);
- else
- match = !fnmatch (str, base_version, 0);
- }
- else /* Substring search */
- {
- if (strstr (version, str))
- match = 1;
- else
- match = 0;
- }
- }
- /* If Wget is not matched, skip all the entries up to the
- next User-agent field. */
- wget_matched = match;
- }
- else if (!wget_matched)
- {
- xfree (line);
- DEBUGP (("(chucking out since it is not applicable for Wget)\n"));
- continue;
- }
- else if (!strcasecmp (cmd, "Disallow"))
- {
- /* If "Disallow" is empty, the robot is welcome. */
- if (!*str)
- {
- free_vec (entries);
- entries = (char **)xmalloc (sizeof (char *));
- *entries = NULL;
- num = 0;
- }
- else
- {
- entries = (char **)xrealloc (entries, (num + 2)* sizeof (char *));
- entries[num] = xstrdup (str);
- entries[++num] = NULL;
- /* Strip trailing spaces, according to specifications. */
- for (i = strlen (str); i >= 0 && ISSPACE (str[i]); i--)
- if (ISSPACE (str[i]))
- str[i] = '\0';
- }
- }
- else
- {
- /* unknown command */
- DEBUGP (("(chucked out)\n"));
- }
- xfree (line);
- }
- fclose (fp);
- return entries;
-}
-
-/* May the URL url be loaded according to disallowing rules stored in
- forbidden? */
-static int
-robots_match (struct urlinfo *u, char **fb)
-{
- int l;
-
- if (!fb)
- return 1;
- DEBUGP (("Matching %s against: ", u->path));
- for (; *fb; fb++)
- {
- DEBUGP (("%s ", *fb));
- l = strlen (*fb);
- /* If dir is fb, we may not load the file. */
- if (strncmp (u->path, *fb, l) == 0)
- {
- DEBUGP (("matched.\n"));
- return 0; /* Matches, i.e. does not load... */
- }
- }
- DEBUGP (("not matched.\n"));
- return 1;
-}
projects / wget / blobdiff