cookies.c

wget讓你可以在console介面下

	return 0;
    }

  DEBUGP ((" 8"));

  return 1;
}

static int path_matches PARAMS ((const char *, const char *));

/* Check whether PATH begins with COOKIE_PATH. */

static int
check_path_match (const char *cookie_path, const char *path)
{
  return path_matches (path, cookie_path);
}

/* Prepend '/' to string S.  S is copied to fresh stack-allocated
   space and its value is modified to point to the new location.  */

#define PREPEND_SLASH(s) do {					\
  char *PS_newstr = (char *) alloca (1 + strlen (s) + 1);	\
  *PS_newstr = '/';						\
  strcpy (PS_newstr + 1, s);					\
  s = PS_newstr;						\
} while (0)


/* Process the HTTP `Set-Cookie' header.  This results in storing the
   cookie or discarding a matching one, or ignoring it completely, all
   depending on the contents.  */

void
cookie_handle_set_cookie (struct cookie_jar *jar,
			  const char *host, int port,
			  const char *path, const char *set_cookie)
{
  struct cookie *cookie;
  cookies_now = time (NULL);

  /* Wget's paths don't begin with '/' (blame rfc1808), but cookie
     usage assumes /-prefixed paths.  Until the rest of Wget is fixed,
     simply prepend slash to PATH.  */
  PREPEND_SLASH (path);

  cookie = parse_set_cookies (set_cookie, update_cookie_field, 0);
  if (!cookie)
    goto out;

  /* Sanitize parts of cookie. */

  if (!cookie->domain)
    {
    copy_domain:
      /* If the domain was not provided, we use the one we're talking
	 to, and set exact match.  */
      cookie->domain = xstrdup (host);
      cookie->domain_exact = 1;
      /* Set the port, but only if it's non-default. */
      if (port != 80 && port != 443)
	cookie->port = port;
    }
  else
    {
      if (!check_domain_match (cookie->domain, host))
	{
	  logprintf (LOG_NOTQUIET,
		     _("Cookie coming from %s attempted to set domain to %s\n"),
		     escnonprint (host), escnonprint (cookie->domain));
	  xfree (cookie->domain);
	  goto copy_domain;
	}
    }

  if (!cookie->path)
    {
      /* The cookie doesn't set path: set it to the URL path, sans the
	 file part ("/dir/file" truncated to "/dir/").  */
      char *trailing_slash = strrchr (path, '/');
      if (trailing_slash)
	cookie->path = strdupdelim (path, trailing_slash + 1);
      else
	/* no slash in the string -- can this even happen? */
	cookie->path = xstrdup (path);
    }
  else
    {
      /* The cookie sets its own path; verify that it is legal. */
      if (!check_path_match (cookie->path, path))
	{
	  DEBUGP (("Attempt to fake the path: %s, %s\n",
		   cookie->path, path));
	  goto out;
	}
    }

  /* Now store the cookie, or discard an existing cookie, if
     discarding was requested.  */

  if (cookie->discard_requested)
    {
      discard_matching_cookie (jar, cookie);
      goto out;
    }

  store_cookie (jar, cookie);
  return;

 out:
  if (cookie)
    delete_cookie (cookie);
}

/* Support for sending out cookies in HTTP requests, based on
   previously stored cookies.  Entry point is
   `build_cookies_request'.  */
   
/* Return a count of how many times CHR occurs in STRING. */

static int
count_char (const char *string, char chr)
{
  const char *p;
  int count = 0;
  for (p = string; *p; p++)
    if (*p == chr)
      ++count;
  return count;
}

/* Find the cookie chains whose domains match HOST and store them to
   DEST.

   A cookie chain is the head of a list of cookies that belong to a
   host/domain.  Given HOST "img.search.xemacs.org", this function
   will return the chains for "img.search.xemacs.org",
   "search.xemacs.org", and "xemacs.org" -- those of them that exist
   (if any), that is.

   DEST should be large enough to accept (in the worst case) as many
   elements as there are domain components of HOST.  */

static int
find_chains_of_host (struct cookie_jar *jar, const char *host,
		     struct cookie *dest[])
{
  int dest_count = 0;
  int passes, passcnt;

  /* Bail out quickly if there are no cookies in the jar.  */
  if (!hash_table_count (jar->chains))
    return 0;

  if (numeric_address_p (host))
    /* If host is an IP address, only check for the exact match. */
    passes = 1;
  else
    /* Otherwise, check all the subdomains except the top-level (last)
       one.  As a domain with N components has N-1 dots, the number of
       passes equals the number of dots.  */
    passes = count_char (host, '.');

  passcnt = 0;

  /* Find chains that match HOST, starting with exact match and
     progressing to less specific domains.  For instance, given HOST
     fly.srk.fer.hr, first look for fly.srk.fer.hr's chain, then
     srk.fer.hr's, then fer.hr's.  */
  while (1)
    {
      struct cookie *chain = hash_table_get (jar->chains, host);
      if (chain)
	dest[dest_count++] = chain;
      if (++passcnt >= passes)
	break;
      host = strchr (host, '.') + 1;
    }

  return dest_count;
}

/* If FULL_PATH begins with PREFIX, return the length of PREFIX, zero
   otherwise.  */

static int
path_matches (const char *full_path, const char *prefix)
{
  int len = strlen (prefix);

  if (0 != strncmp (full_path, prefix, len))
    /* FULL_PATH doesn't begin with PREFIX. */
    return 0;

  /* Length of PREFIX determines the quality of the match. */
  return len + 1;
}

/* Return non-zero iff COOKIE matches the provided parameters of the
   URL being downloaded: HOST, PORT, PATH, and SECFLAG.

   If PATH_GOODNESS is non-NULL, store the "path goodness" value
   there.  That value is a measure of how closely COOKIE matches PATH,
   used for ordering cookies.  */

static int
cookie_matches_url (const struct cookie *cookie,
		    const char *host, int port, const char *path,
		    int secflag, int *path_goodness)
{
  int pg;

  if (cookie_expired_p (cookie))
    /* Ignore stale cookies.  Don't bother unchaining the cookie at
       this point -- Wget is a relatively short-lived application, and
       stale cookies will not be saved by `save_cookies'.  On the
       other hand, this function should be as efficient as
       possible.  */
    return 0;

  if (cookie->secure && !secflag)
    /* Don't transmit secure cookies over insecure connections.  */
    return 0;
  if (cookie->port != PORT_ANY && cookie->port != port)
    return 0;

  /* If exact domain match is required, verify that cookie's domain is
     equal to HOST.  If not, assume success on the grounds of the
     cookie's chain having been found by find_chains_of_host.  */
  if (cookie->domain_exact
      && 0 != strcasecmp (host, cookie->domain))
    return 0;

  pg = path_matches (path, cookie->path);
  if (!pg)
    return 0;

  if (path_goodness)
    /* If the caller requested path_goodness, we return it.  This is
       an optimization, so that the caller doesn't need to call
       path_matches() again.  */
    *path_goodness = pg;
  return 1;
}

/* A structure that points to a cookie, along with the additional
   information about the cookie's "goodness".  This allows us to sort
   the cookies when returning them to the server, as required by the
   spec.  */

struct weighed_cookie {
  struct cookie *cookie;
  int domain_goodness;
  int path_goodness;
};

/* Comparator used for uniquifying the list. */

static int
equality_comparator (const void *p1, const void *p2)
{
  struct weighed_cookie *wc1 = (struct weighed_cookie *)p1;
  struct weighed_cookie *wc2 = (struct weighed_cookie *)p2;

  int namecmp  = strcmp (wc1->cookie->attr, wc2->cookie->attr);
  int valuecmp = strcmp (wc1->cookie->value, wc2->cookie->value);

  /* We only really care whether both name and value are equal.  We
     return them in this order only for consistency...  */
  return namecmp ? namecmp : valuecmp;
}

/* Eliminate duplicate cookies.  "Duplicate cookies" are any two
   cookies with the same attr name and value.  Whenever a duplicate
   pair is found, one of the cookies is removed.  */

static int
eliminate_dups (struct weighed_cookie *outgoing, int count)
{
  struct weighed_cookie *h;	/* hare */
  struct weighed_cookie *t;	/* tortoise */
  struct weighed_cookie *end = outgoing + count;

  /* We deploy a simple uniquify algorithm: first sort the array
     according to our sort criteria, then copy it to itself, comparing
     each cookie to its neighbor and ignoring the duplicates.  */

  qsort (outgoing, count, sizeof (struct weighed_cookie), equality_comparator);

  /* "Hare" runs through all the entries in the array, followed by
     "tortoise".  If a duplicate is found, the hare skips it.
     Non-duplicate entries are copied to the tortoise ptr.  */

  for (h = t = outgoing; h < end; h++)
    {
      if (h != end - 1)
	{
	  struct cookie *c0 = h[0].cookie;
	  struct cookie *c1 = h[1].cookie;
	  if (!strcmp (c0->attr, c1->attr) && !strcmp (c0->value, c1->value))
	    continue;		/* ignore the duplicate */
	}

      /* If the hare has advanced past the tortoise (because of
	 previous dups), make sure the values get copied.  Otherwise,
	 no copying is necessary.  */
      if (h != t)
	*t++ = *h;
      else
	t++;
    }
  return t - outgoing;
}

/* Comparator used for sorting by quality. */

static int
goodness_comparator (const void *p1, const void *p2)
{
  struct weighed_cookie *wc1 = (struct weighed_cookie *)p1;
  struct weighed_cookie *wc2 = (struct weighed_cookie *)p2;

  /* Subtractions take `wc2' as the first argument becauase we want a
     sort in *decreasing* order of goodness.  */
  int dgdiff = wc2->domain_goodness - wc1->domain_goodness;
  int pgdiff = wc2->path_goodness - wc1->path_goodness;

  /* Sort by domain goodness; if these are the same, sort by path
     goodness.  (The sorting order isn't really specified; maybe it
     should be the other way around.)  */
  return dgdiff ? dgdiff : pgdiff;
}

/* Generate a `Cookie' header for a request that goes to HOST:PORT and
   requests PATH from the server.  The resulting string is allocated
   with `malloc', and the caller is responsible for freeing it.  If no
   cookies pertain to this request, i.e. no cookie header should be
   generated, NULL is returned.  */

char *
cookie_header (struct cookie_jar *jar, const char *host,
	       int port, const char *path, int secflag)
{
  struct cookie **chains;
  int chain_count;

  struct cookie *cookie;
  struct weighed_cookie *outgoing;
  int count, i, ocnt;
  char *result;
  int result_size, pos;
  PREPEND_SLASH (path);		/* see cookie_handle_set_cookie */

  /* First, find the cookie chains whose domains match HOST. */

  /* Allocate room for find_chains_of_host to write to.  The number of
     chains can at most equal the number of subdomains, hence
     1+<number of dots>.  */
  chains = alloca_array (struct cookie *, 1 + count_char (host, '.'));
  chain_count = find_chains_of_host (jar, host, chains);

  /* No cookies for this host. */
  if (!chain_count)
    return NULL;

  cookies_now = time (NULL);

  /* Now extract from the chains those cookies that match our host
     (for domain_exact cookies), port (for cookies with port other
     than PORT_ANY), etc.  See matching_cookie for details.  */

  /* Count the number of matching cookies. */
  count = 0;
  for (i = 0; i < chain_count; i++)
    for (cookie = chains[i]; cookie; cookie = cookie->next)
      if (cookie_matches_url (cookie, host, port, path, secflag, NULL))
	++count;
  if (!count)
    return NULL;		/* no cookies matched */

  /* Allocate the array. */
  outgoing = alloca_array (struct weighed_cookie, count);

  /* Fill the array with all the matching cookies from the chains that
     match HOST. */
  ocnt = 0;
  for (i = 0; i < chain_count; i++)
    for (cookie = chains[i]; cookie; cookie = cookie->next)
      {
	int pg;
	if (!cookie_matches_url (cookie, host, port, path, secflag, &pg))
	  continue;
	outgoing[ocnt].cookie = cookie;
	outgoing[ocnt].domain_goodness = strlen (cookie->domain);
	outgoing[ocnt].path_goodness   = pg;
	++ocnt;
      }
  assert (ocnt == count);

  /* Eliminate duplicate cookies; that is, those whose name and value
     are the same.  */
  count = eliminate_dups (outgoing, count);

  /* Sort the array so that best-matching domains come first, and
     that, within one domain, best-matching paths come first. */
  qsort (outgoing, count, sizeof (struct weighed_cookie), goodness_comparator);