Merge pull request #128 from azmeuk/self-closing

Replaced XHTML output with HTML5 output in summary for empty elements (a, br), issue #125
4 years ago · da9e285f73
parent 5032e2d3ab 0846955dd7
commit da9e285f73
2 changed files with 6 additions and 1 deletions
--- a/readability/readability.py
+++ b/readability/readability.py
@ -171,7 +171,7 @@ class Document:
        An internal method, which can be overridden in subclasses, for example,
        to disable or to improve DOM-to-text conversion in .summary() method
        """
-        return clean_attributes(tounicode(self.html))
+        return clean_attributes(tounicode(self.html, method='html'))

    def summary(self, html_partial=False):
        """
--- a/tests/test_article_only.py
+++ b/tests/test_article_only.py
@ -106,3 +106,8 @@ class TestArticleOnly(unittest.TestCase):
        s = doc.summary()

        assert 'foo' in s
+
+    def test_not_self_closing(self):
+        sample = '<h2><a href="#"></a>foobar</h2>'
+        doc = Document(sample)
+        assert '<body id="readabilityBody"><h2><a href="#"></a>foobar</h2></body>' == doc.summary()