Fix HTML handling of </>

docs/changelog.md

@@ -14,6 +14,7 @@ See the [Contributing Guide](contributing.md) for details.
 ### Fixed
 
 * Fixed dropped content in `md_in_html` (#1526).
+* Fixed HTML handling corner case that prevented some content from not being rendered (#1528).
 
 ## [3.8.0] - 2025-04-09
 

markdown/htmlparser.py

@@ -41,6 +41,10 @@
 spec.loader.exec_module(htmlparser)
 sys.modules['htmlparser'] = htmlparser
 
+# This is a hack. We are sneaking in `</>` so we can capture it without the HTML parser
+# throwing it away. When we see it, we will process it as data.
+htmlparser.starttagopen = re.compile('<[a-zA-Z]|</>')
+
 # Monkeypatch `HTMLParser` to only accept `?>` to close Processing Instructions.
 htmlparser.piclose = re.compile(r'\?>')
 # Monkeypatch `HTMLParser` to only recognize entity references with a closing semicolon.
@@ -297,6 +301,11 @@ def get_starttag_text(self) -> str:
         return self.__starttag_text
 
     def parse_starttag(self, i: int) -> int:  # pragma: no cover
+        # Treat `</>` as normal data as it is not a real tag.
+        if self.rawdata[i:i + 3] == '</>':
+            self.handle_data(self.rawdata[i:i + 3])
+            return i + 3
+
         self.__starttag_text = None
         endpos = self.check_for_whole_start_tag(i)
         if endpos < 0:

tests/test_syntax/blocks/test_html_blocks.py

@@ -1643,3 +1643,21 @@ def test_placeholder_in_source(self):
         placeholder = md.htmlStash.get_placeholder(md.htmlStash.html_counter + 1)
         result = md.postprocessors['raw_html'].run(placeholder)
         self.assertEqual(placeholder, result)
+
+    def test_noname_tag(self):
+        self.assertMarkdownRenders(
+            self.dedent(
+                """
+                <div>
+                </>
+                </div>
+                """
+            ),
+            self.dedent(
+                """
+                <div>
+                </>
+                </div>
+                """
+            )
+        )

tests/test_syntax/extensions/test_md_in_html.py

@@ -1538,6 +1538,24 @@ def test_trailing_content_after_tag_in_md_block(self):
             extensions=['md_in_html']
         )
 
+    def test_noname_tag(self):
+        self.assertMarkdownRenders(
+            self.dedent(
+                """
+                <div markdown>
+                </>
+                </div>
+                """
+            ),
+            self.dedent(
+                """
+                <div>
+                <p>&lt;/&gt;</p>
+                </div>
+                """
+            )
+        )
+
 
 def load_tests(loader, tests, pattern):
     """ Ensure `TestHTMLBlocks` doesn't get run twice by excluding it here. """

tests/test_syntax/inline/test_code.py

@@ -62,3 +62,20 @@ def test_code_html(self):
                 """
             )
         )
+
+    def test_noname_tag(self):
+        # Browsers ignore `</>`, but a Markdown parser should not, and should treat it as data
+        # but not a tag.
+
+        self.assertMarkdownRenders(
+            self.dedent(
+                """
+                `</>`
+                """
+            ),
+            self.dedent(
+                """
+                <p><code>&lt;/&gt;</code></p>
+                """
+            )
+        )

tests/test_syntax/inline/test_raw_html.py

@@ -31,3 +31,6 @@ def test_inline_html_angle_brackets(self):
 
     def test_inline_html_backslashes(self):
         self.assertMarkdownRenders('<img src="..\\..\\foo.png">', '<p><img src="..\\..\\foo.png"></p>')
+
+    def test_noname_tag(self):
+        self.assertMarkdownRenders('<span></></span>', '<p><span>&lt;/&gt;</span></p>')