[fix] handle single closing element in HTMLTextExtractor

2015-01-22 17:43:45 +01:00 · 2015-01-22 17:43:45 +01:00 · 699fe60e50
commit 699fe60e50
parent 06186e72a9
1 changed files with 4 additions and 0 deletions
--- a/searx/utils.py
+++ b/searx/utils.py
@ -82,8 +82,12 @@ class HTMLTextExtractor(HTMLParser):
        self.tags.append(tag)

    def handle_endtag(self, tag):
+        if not self.tags:
+            return
+
        if tag != self.tags[-1]:
            raise Exception("invalid html")
+
        self.tags.pop()

    def is_valid_tag(self):