From: Victor Stinner <victor.stinner@haypocalc.com>
Date: Mon, 24 May 2010 21:33:24 +0000 (+0000)
Subject: Issue #6662: Fix parsing of malformatted charref (&#bad;)
X-Git-Tag: v2.7rc1~95
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=554a3b82e40573846f893ffdfff230e1d908af57;p=python

Issue #6662: Fix parsing of malformatted charref (&#bad;)
---

diff --git a/Lib/HTMLParser.py b/Lib/HTMLParser.py
index 2cbc2ecbc7..7cee47a7c5 100644
--- a/Lib/HTMLParser.py
+++ b/Lib/HTMLParser.py
@@ -175,6 +175,9 @@ class HTMLParser(markupbase.ParserBase):
                     i = self.updatepos(i, k)
                     continue
                 else:
+                    if ";" in rawdata[i:]: #bail by consuming &#
+                        self.handle_data(rawdata[0:2])
+                        i = self.updatepos(i, 2)
                     break
             elif startswith('&', i):
                 match = entityref.match(rawdata, i)
diff --git a/Lib/test/test_htmlparser.py b/Lib/test/test_htmlparser.py
index 810af6c8cb..c45cf00ece 100644
--- a/Lib/test/test_htmlparser.py
+++ b/Lib/test/test_htmlparser.py
@@ -313,6 +313,13 @@ DOCTYPE html [
                 ("starttag", "html", [("foo", u"\u20AC&aa&unsupported;")])
                 ])
 
+    def test_malformatted_charref(self):
+        self._run_check("<p>&#bad;</p>", [
+            ("starttag", "p", []),
+            ("data", "&#bad;"),
+            ("endtag", "p"),
+        ])
+
 
 def test_main():
     test_support.run_unittest(HTMLParserTestCase)
diff --git a/Misc/NEWS b/Misc/NEWS
index b094126889..e2f0f0ccf8 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -29,6 +29,8 @@ C-API
 Library
 -------
 
+- Issue #6662: Fix parsing of malformatted charref (&#bad;)
+
 - Issue #8016: Add the CP858 codec.
 
 - Issue #3924: Ignore cookies with invalid "version" field in cookielib.