New function 'domain_get_tld()'

This function  can identify attemtps to have domains
above the allowed TLD - which can then be used to identify
supercookies et.al.

1 files changed, 51 insertions, 7 deletions

diff --git a/src/domain.c b/src/domain.c
index 432396e1..9aa6f156 100644
--- a/src/domain.c
+++ b/src/domain.c
@@ -1,5 +1,6 @@
 /*
  * Copyright (c) 2010-2013 Stefan Bolte <portix@gmx.net>
+ * Copyright (c) 2013 Elias Norberg <xyzzy@kudzu.se>
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -124,8 +125,22 @@ domain_match(char **domains, const char *host, const char *base_domain)
     return false;
 }/*}}}*/
 
+int
+count_char(const char *str, char ch)
+{
+    int count = 0;
+
+    if (str == NULL) 
+        return 0;
+
+    for (; *str != '\0'; str++)
+        if (*str == ch)
+            count ++;
+    return count;
+}
+
 const char *
-domain_get_base_for_host(const char *host) 
+domain_get_tld(const char *host)
 {
     if (host == NULL)
         return NULL;
@@ -138,6 +153,17 @@ domain_get_base_for_host(const char *host)
     char *nextdot = strchr(cur_domain, '.');
     char *entry = NULL;
 
+    /* check if hostname is valid
+     * - cannot start with .
+     * - must only contain A-Za-z0-9.-_
+     */
+    if (nextdot == cur_domain ||
+        strspn(cur_domain, "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+                    "abcdefghijklmnopqrustuvwxyz"
+                    "0123456789._-") != strlen(cur_domain)) {
+        return NULL;
+    }
+
     while (1) 
     {
         entry = g_hash_table_lookup(s_tld_table, cur_domain);
@@ -145,15 +171,23 @@ domain_get_base_for_host(const char *host)
         {
             if (*entry == '*') 
             {
-                ret = pprev_domain;
+                /* check that we have enough to build
+                 * a domain that satisfies the rule
+                 */
+                if (count_char(pprev_domain, '.') >=
+                    count_char(entry, '.') + 1) {
+                    ret = pprev_domain;
+                }
                 break;
             }
-            else if (*entry == '!' && nextdot) 
+            else if (*entry == '!') 
             {
-                ret = nextdot + 1;
+                ret = cur_domain;
                 break;
             }
-            else 
+            else if(strcmp(entry, prev_domain) == 0)
+                break;
+            else if(strcmp(entry, prev_domain) != 0)
             {
                 ret = prev_domain;
                 break;
@@ -170,10 +204,20 @@ domain_get_base_for_host(const char *host)
 
         nextdot = strchr(cur_domain, '.');
     }
-    if (ret == NULL) 
-        ret = host;
     return ret;
 }
+
+
+const char *
+domain_get_base_for_host(const char *host) 
+{
+    const char *base;
+    base = domain_get_tld(host);
+    if (base == NULL)
+        return host;
+    return base;
+}
+
 void
 domain_end() 
 {