#20288: fix handling of invalid numeric charrefs in HTMLParser.

Lib/html/parser.py

@@ -228,9 +228,9 @@ def goahead(self, end):
                     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)
+                    if ";" in rawdata[i:]:  # bail by consuming &#
+                        self.handle_data(rawdata[i:i+2])
+                        i = self.updatepos(i, i+2)
                     break
             elif startswith('&', i):
                 match = entityref.match(rawdata, i)

Lib/test/test_htmlparser.py

@@ -151,6 +151,12 @@ def test_malformatted_charref(self):
             ("data", "&#bad;"),
             ("endtag", "p"),
         ])
+        # add the [] as a workaround to avoid buffering (see #20288)
+        self._run_check(["<div>&#bad;</div>"], [
+            ("starttag", "div", []),
+            ("data", "&#bad;"),
+            ("endtag", "div"),
+        ])
 
     def test_unclosed_entityref(self):
         self._run_check("&entityref foo", [

Misc/NEWS

@@ -45,6 +45,8 @@ Core and Builtins
 Library
 -------
 
+- Issue #20288: fix handling of invalid numeric charrefs in HTMLParser.
+
 - Issue #20424: Python implementation of io.StringIO now supports lone surrogates.
 
 - Issue #19456: ntpath.join() now joins relative paths correctly when a drive