* Fixes to fuzzy hashing logic, skip urls while estimating fuzzy hash

Fix tags stripping.
Fix phishing checks (ignore img tags).

2 files changed, 21 insertions, 25 deletions

diff --git a/src/plugins/fuzzy_check.c b/src/plugins/fuzzy_check.c
index 3f068ed99..93ba4bf25 100644
--- a/src/plugins/fuzzy_check.c
+++ b/src/plugins/fuzzy_check.c
@@ -686,6 +686,7 @@ fuzzy_symbol_callback (struct worker_task *task, void *unused)
 		}
 
 		register_fuzzy_call (task, part->fuzzy);
+		register_fuzzy_call (task, part->double_fuzzy);
 
 		cur = g_list_next (cur);
 	}
@@ -843,6 +844,16 @@ fuzzy_process_handler (struct controller_session *session, f_str_t * in)
 				free_task (task, FALSE);
 				return;
 			}
+			if (! register_fuzzy_controller_call (session, task, part->double_fuzzy, cmd, value, flag, saved)) {
+				/* Cannot write hash */
+				session->state = STATE_REPLY;
+				r = rspamd_snprintf (out_buf, sizeof (out_buf), "cannot write fuzzy hash" CRLF "END" CRLF);
+				if (! rspamd_dispatcher_write (session->dispatcher, out_buf, r, FALSE, FALSE)) {
+					return;
+				}
+				free_task (task, FALSE);
+				return;
+			}
 			cur = g_list_next (cur);
 		}
 		/* Process images */
diff --git a/src/plugins/regexp.c b/src/plugins/regexp.c
index 441a17de5..a82c60110 100644
--- a/src/plugins/regexp.c
+++ b/src/plugins/regexp.c
@@ -845,31 +845,16 @@ process_regexp (struct rspamd_regexp *re, struct worker_task *task, const gchar
 			/*XXX: add support of it */
 			msg_warn ("numbered matches are not supported for url regexp");
 		}
-		cur = g_list_first (task->text_parts);
-		while (cur) {
-			part = (struct mime_text_part *)cur->data;
-			/* Skip empty parts */
-			if (part->is_empty) {
-				cur = g_list_next (cur);
-				continue;
-			}
-			if (part->is_raw) {
-				regexp = re->raw_regexp;
-			}
-			else {
-				regexp = re->regexp;
-			}
-			callback_param.task = task;
-			callback_param.regexp = regexp;
-			callback_param.re = re;
-			callback_param.found = FALSE;
-			if (part->urls) {
-				g_tree_foreach (part->urls, tree_url_callback, &callback_param);
-			}
-			if (part->html_urls && callback_param.found == FALSE) {
-				g_tree_foreach (part->html_urls, tree_url_callback, &callback_param);
-			}
-			cur = g_list_next (cur);
+		regexp = re->regexp;
+		callback_param.task = task;
+		callback_param.regexp = regexp;
+		callback_param.re = re;
+		callback_param.found = FALSE;
+		if (task->urls) {
+			g_tree_foreach (task->urls, tree_url_callback, &callback_param);
+		}
+		if (task->emails && callback_param.found == FALSE) {
+			g_tree_foreach (task->emails, tree_url_callback, &callback_param);
 		}
 		if (callback_param.found == FALSE) {
 			task_cache_add (task, re, 0);