Fix phishing detection with img flag.

Handle unclosed HTML tags properly.
Remove warnings for types on 32 bit archs.
Do not touch grow factor many times when one shot mode is turned on.

diff --git a/src/dns.c b/src/dns.c
index 4adab9af90af16e0b42283a7d72397f3c0b0a820..e82b1a9b723de88a604f9d5a560085d6b591fe82 100644 (file)
--- a/src/dns.c
+++ b/src/dns.c
@@ -566,7 +566,7 @@ dns_fin_cb (gpointer arg)
        struct rspamd_dns_request *req = arg;
        
        event_del (&req->timer_event);
-       g_hash_table_remove (req->resolver->requests, GUINT_TO_POINTER (req->id));
+       g_hash_table_remove (req->resolver->requests, GUINT_TO_POINTER ((guint)req->id));
 }
 
 static guint8 *
@@ -925,7 +925,7 @@ dns_parse_reply (guint8 *in, gint r, struct rspamd_dns_resolver *resolver,
        }
 
        /* Now try to find corresponding request */
-       if ((req = g_hash_table_lookup (resolver->requests, GUINT_TO_POINTER (header->qid))) == NULL) {
+       if ((req = g_hash_table_lookup (resolver->requests, GUINT_TO_POINTER ((guint)header->qid))) == NULL) {
                /* No such requests found */
                return FALSE;
        }
@@ -1134,7 +1134,7 @@ dns_retransmit_handler (gint fd, short what, void *arg)
                        evtimer_add (&req->timer_event, &req->tv);
 
                        /* Add request to hash table */
-                       g_hash_table_insert (req->resolver->requests, GUINT_TO_POINTER (req->id), req); 
+                       g_hash_table_insert (req->resolver->requests, GUINT_TO_POINTER ((guint)req->id), req);
                        register_async_event (req->session, (event_finalizer_t)dns_fin_cb, req, FALSE);
                }
        }
@@ -1233,7 +1233,7 @@ make_dns_request (struct rspamd_dns_resolver *resolver,
                evtimer_add (&req->timer_event, &req->tv);
 
                /* Add request to hash table */
-               g_hash_table_insert (resolver->requests, GUINT_TO_POINTER (req->id), req); 
+               g_hash_table_insert (resolver->requests, GUINT_TO_POINTER ((guint)req->id), req);
                register_async_event (session, (event_finalizer_t)dns_fin_cb, req, FALSE);
        }
        else if (r == -1) {

diff --git a/src/filter.c b/src/filter.c
index b48bf64ba25c83d7d5b3fe94423fb57388bf4445..fea91125fee695b9f1c1528f3c5035596532ea11 100644 (file)
--- a/src/filter.c
+++ b/src/filter.c
@@ -80,18 +80,9 @@ insert_metric_result (struct worker_task *task, struct metric *metric, const gch
        else {
                w = (*weight) * flag;
        }
-       /* Handle grow factor */
-       if (metric_res->grow_factor && w > 0) {
-               w *= metric_res->grow_factor;
-               metric_res->grow_factor *= metric->grow_factor;
-       }
-       else if (w > 0) {
-               metric_res->grow_factor = metric->grow_factor;
-       }
-
-       /* Add metric score */
 
 
+       /* Add metric score */
        if ((s = g_hash_table_lookup (metric_res->symbols, symbol)) != NULL) {
                if (s->options && opts && opts != s->options) {
                        /* Append new options */
@@ -106,6 +97,14 @@ insert_metric_result (struct worker_task *task, struct metric *metric, const gch
                        memory_pool_add_destructor (task->task_pool, (pool_destruct_func) g_list_free, s->options);
                }
                if (!single) {
+                       /* Handle grow factor */
+                       if (metric_res->grow_factor && w > 0) {
+                               w *= metric_res->grow_factor;
+                               metric_res->grow_factor *= metric->grow_factor;
+                       }
+                       else if (w > 0) {
+                               metric_res->grow_factor = metric->grow_factor;
+                       }
                        s->score += w;
                        metric_res->score += w;
                }
@@ -114,6 +113,14 @@ insert_metric_result (struct worker_task *task, struct metric *metric, const gch
                s = memory_pool_alloc (task->task_pool, sizeof (struct symbol));
                s->score = w;
 
+               /* Handle grow factor */
+               if (metric_res->grow_factor && w > 0) {
+                       w *= metric_res->grow_factor;
+                       metric_res->grow_factor *= metric->grow_factor;
+               }
+               else if (w > 0) {
+                       metric_res->grow_factor = metric->grow_factor;
+               }
                s->name = symbol;
                metric_res->score += w;
 

diff --git a/src/html.c b/src/html.c
index 3582022f8db34d286d7db7cab175a3853cba3274..bf261082140d1211f95ca5fe9c11a71a25aa4118 100644 (file)
--- a/src/html.c
+++ b/src/html.c
@@ -662,7 +662,7 @@ static gchar *
 html_strncasestr (const gchar *s, const gchar *find, gsize len)
 {
        gchar                           c, sc;
-       size_t mlen;
+       gsize                           mlen;
 
        if ((c = *find++) != 0) {
                c = g_ascii_tolower (c);
@@ -692,26 +692,41 @@ check_phishing (struct worker_task *task, struct uri *href_url, const gchar *url
        p = url_text;
        while (len < remain) {
                if (*p == '<') {
-                       /* Get tag name */
-                       p ++;
-                       len ++;
-                       if (*p == '/') {
-                               /* Check tag name */
+                       /* Check tag name */
+                       if (*(p + 1) == '/') {
+                               c = p + 2;
+                       }
+                       else {
                                c = p + 1;
-                               while (len < remain) {
-                                       if (!g_ascii_isspace (*p) && *p != '>') {
+                       }
+                       while (len < remain) {
+                               if (!g_ascii_isspace (*p) && *p != '>') {
+                                       p ++;
+                                       len ++;
+                               }
+                               else {
+                                       break;
+                               }
+                       }
+                       rspamd_strlcpy (tagbuf, c, MIN (sizeof(tagbuf), p - c + 1));
+                       if ((tag = get_tag_by_name (tagbuf)) != NULL) {
+                               if (tag->id == id) {
+                                       break;
+                               }
+                               else if (tag->id == Tag_IMG) {
+                                       /* We should ignore IMG tag here */
+                                       while (len < remain && *p != '>' && *p != '<') {
                                                p ++;
                                                len ++;
                                        }
-                                       else {
-                                               break;
-                                       }
-                               }
-                               rspamd_strlcpy (tagbuf, c, MIN (sizeof(tagbuf), p - c + 1));
-                               if ((tag = get_tag_by_name (tagbuf)) != NULL) {
-                                       if (tag->id == id) {
-                                               break;
+                                       if (*p == '>' && len < remain) {
+                                               p ++;
                                        }
+
+                                       remain -= p - url_text;
+                                       url_text = p;
+                                       len = 0;
+                                       continue;
                                }
                        }
                }
@@ -724,6 +739,7 @@ check_phishing (struct worker_task *task, struct uri *href_url, const gchar *url
                if (new != NULL) {
                        g_strstrip (url_str);
                        rc = parse_uri (new, url_str, task->task_pool);
+
                        if (rc == URI_ERRNO_OK || rc == URI_ERRNO_NO_SLASHES || rc == URI_ERRNO_NO_HOST_SLASH) {
                                if (g_ascii_strncasecmp (href_url->host, new->host,
                                                MAX (href_url->hostlen, new->hostlen)) != 0) {

diff --git a/src/message.c b/src/message.c
index 4db4bef7df4d7c716da5f6081d95014c5c3a21b9..8ff53ea938a6ef1c410ab9295d474d4257a2eaa6 100644 (file)
--- a/src/message.c
+++ b/src/message.c
@@ -70,7 +70,13 @@ strip_html_tags (struct worker_task *task, memory_pool_t * pool, struct mime_tex
                                state = 1;
                        }
                        else if (state == 1) {
-                               depth++;
+                               /* Opening bracket without closing one */
+                               p --;
+                               while (g_ascii_isspace (*p) && p > src->data) {
+                                       p --;
+                               }
+                               p ++;
+                               goto unbreak_tag;
                        }
                        break;
 
@@ -107,7 +113,7 @@ strip_html_tags (struct worker_task *task, memory_pool_t * pool, struct mime_tex
                        if (in_q) {
                                break;
                        }
-
+unbreak_tag:
                        switch (state) {
                        case 1:                 /* HTML/XML */
                                lc = '>';