Improve the robustness of the HTML parser

1 files changed, 9 insertions, 0 deletions

diff --git a/libmat2/web.py b/libmat2/web.py
index 2864d60..8843c1e 100644
--- a/libmat2/web.py
+++ b/libmat2/web.py
@@ -104,6 +104,15 @@ class _HTMLParser(parser.HTMLParser):
         self.tag_required_blocklist = required_blocklisted_tags
         self.tag_blocklist = blocklisted_tags
 
+    def error(self, message):  # pragma: no cover
+        """ Amusingly, Python's documentation doesn't mention that this
+        function needs to be implemented in subclasses of the parent class
+        of parser.HTMLParser. This was found by fuzzing,
+        triggering the following exception:
+            NotImplementedError: subclasses of ParserBase must override error()
+        """
+        raise ValueError(message)
+
     def handle_starttag(self, tag: str, attrs: List[Tuple[str, Optional[str]]]):
         # Ignore the type, because mypy is too stupid to infer
         # that get_starttag_text() can't return None.