[svn] Use the new function `random_number' that doesn't depend on RAND_MAX

[wget] / src / retr.c

diff --git a/src/retr.c b/src/retr.c
index 94fa97b2d76e1d87bbec6fd964a044a7cec40f6c..6bd885894c429a0ce7f86897541ba51d1701ed42 100644 (file)
--- a/src/retr.c
+++ b/src/retr.c
@@ -44,6 +44,10 @@ Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.  */
 #include "connect.h"
 #include "hash.h"
 
+#ifdef HAVE_SSL
+# include "gen_sslfunc.h"      /* for ssl_iread */
+#endif
+
 #ifndef errno
 extern int errno;
 #endif
@@ -294,9 +298,16 @@ register_all_redirections (struct hash_table *redirections, const char *final)
                        && no_proxy_match((u)->host,                    \
                                          (const char **)opt.no_proxy))
 
-/* Retrieve the given URL.  Decides which loop to call -- HTTP(S), FTP,
-   or simply copy it with file:// (#### the latter not yet
-   implemented!).  */
+/* Maximum number of allowed redirections.  20 was chosen as a
+   "reasonable" value, which is low enough to not cause havoc, yet
+   high enough to guarantee that normal retrievals will not be hurt by
+   the check.  */
+
+#define MAX_REDIRECTIONS 20
+
+/* Retrieve the given URL.  Decides which loop to call -- HTTP, FTP,
+   FTP, proxy, etc.  */
+
 uerr_t
 retrieve_url (const char *origurl, char **file, char **newloc,
              const char *refurl, int *dt)
@@ -310,6 +321,7 @@ retrieve_url (const char *origurl, char **file, char **newloc,
   int up_error_code;           /* url parse error code */
   char *local_file;
   struct hash_table *redirections = NULL;
+  int redirection_count = 0;
 
   /* If dt is NULL, just ignore it.  */
   if (!dt)
@@ -398,18 +410,17 @@ retrieve_url (const char *origurl, char **file, char **newloc,
        opt.recursive = 0;
       result = ftp_loop (u, dt);
       opt.recursive = oldrec;
-#if 0
+
       /* There is a possibility of having HTTP being redirected to
         FTP.  In these cases we must decide whether the text is HTML
         according to the suffix.  The HTML suffixes are `.html' and
         `.htm', case-insensitive.  */
-      if (redirections && u->local && (u->scheme == SCHEME_FTP))
+      if (redirections && local_file && u->scheme == SCHEME_FTP)
        {
-         char *suf = suffix (u->local);
+         char *suf = suffix (local_file);
          if (suf && (!strcasecmp (suf, "html") || !strcasecmp (suf, "htm")))
            *dt |= TEXTHTML;
        }
-#endif
     }
   location_changed = (result == NEWLOCATION);
   if (location_changed)
@@ -458,7 +469,22 @@ retrieve_url (const char *origurl, char **file, char **newloc,
          string_set_add (redirections, u->url);
        }
 
-      /* The new location is OK.  Check for redirection cycle by
+      /* The new location is OK.  Check for max. number of
+        redirections.  */
+      if (++redirection_count > MAX_REDIRECTIONS)
+       {
+         logprintf (LOG_NOTQUIET, _("%d redirections exceeded.\n"),
+                    MAX_REDIRECTIONS);
+         url_free (newloc_parsed);
+         url_free (u);
+         if (redirections)
+           string_set_free (redirections);
+         xfree (url);
+         xfree (mynewloc);
+         return WRONGCODE;
+       }
+
+      /*Check for redirection cycle by
          peeking through the history of redirections. */
       if (string_set_contains (redirections, newloc_parsed->url))
        {
@@ -485,11 +511,11 @@ retrieve_url (const char *origurl, char **file, char **newloc,
     {
       if (*dt & RETROKF)
        {
-         register_download (url, local_file);
+         register_download (u->url, local_file);
          if (redirections)
-           register_all_redirections (redirections, url);
+           register_all_redirections (redirections, u->url);
          if (*dt & TEXTHTML)
-           register_html (url, local_file);
+           register_html (u->url, local_file);
        }
     }
 
@@ -531,7 +557,7 @@ retrieve_from_file (const char *file, int html, int *count)
   uerr_t status;
   struct urlpos *url_list, *cur_url;
 
-  url_list = (html ? get_urls_html (file, NULL, FALSE, NULL)
+  url_list = (html ? get_urls_html (file, NULL, NULL)
              : get_urls_file (file));
   status = RETROK;             /* Suppose everything is OK.  */
   *count = 0;                  /* Reset the URL count.  */
@@ -627,10 +653,6 @@ sleep_between_retrievals (int count)
 {
   static int first_retrieval = 1;
 
-  if (first_retrieval && opt.random_wait)
-    /* --random-wait uses the RNG, so seed it. */
-    srand (time (NULL));
-
   if (!first_retrieval && (opt.wait || opt.waitretry))
     {
       if (opt.waitretry && count > 1)
@@ -650,10 +672,10 @@ sleep_between_retrievals (int count)
            sleep (opt.wait);
          else
            {
-             int waitmax = 2 * opt.wait;
-             /* This is equivalent to rand() % waitmax, but uses the
-                high-order bits for better randomness.  */
-             int waitsecs = (double)waitmax * rand () / (RAND_MAX + 1.0);
+             /* Sleep a random amount of time averaging in opt.wait
+                seconds.  The sleeping amount ranges from 0 to
+                opt.wait*2, inclusive.  */
+             int waitsecs = random_number (opt.wait * 2 + 1);
 
              DEBUGP (("sleep_between_retrievals: norm=%ld,fuzz=%ld,sleep=%d\n",
                       opt.wait, waitsecs - opt.wait, waitsecs));