rewrite: ensure lxml parser closes gracefully on no input

pywb/rewrite/lxml_html_rewriter.py

@@ -37,13 +37,17 @@ class LXMLHTMLRewriter(HTMLRewriterMixin):
                                             recover=True,
                                             )
 
+        self.started = False
+
     def feed(self, string):
+        self.started = True
         string = self.END_HTML.sub(u'', string)
         #string = string.replace(u'</html>', u'')
         self.parser.feed(string)
 
     def _internal_close(self):
-        self.parser.close()
+        if self.started:
+            self.parser.close()
 
 
 #=================================================================

pywb/rewrite/test/test_html_rewriter.py

@@ -104,6 +104,16 @@ ur"""
 
 >>> parse('<!-- <a href="http://example.com"></a> -->')
 <!-- <a href="http://example.com"></a> -->
+
+# Test blank
+>>> parse('')
+<BLANKLINE>
+
+# Test no parsing at all
+>>> p = HTMLRewriter(urlrewriter)
+>>> p.close()
+''
+
 """
 
 from pywb.rewrite.url_rewriter import UrlRewriter

pywb/rewrite/test/test_lxml_html_rewriter.py

@@ -110,6 +110,15 @@ ur"""
 
 >>> parse('<!-- <a href="http://example.com"></a> -->')
 <!-- <a href="http://example.com"></a> -->
+
+# Test blank
+>>> parse('')
+<BLANKLINE>
+
+# Test no parsing at all
+>>> p = LXMLHTMLRewriter(urlrewriter)
+>>> p.close()
+''
 """
 
 from pywb.rewrite.url_rewriter import UrlRewriter