Avoid rescanning for a comment close after unclosed comments

.spell-dict

@@ -108,6 +108,7 @@ refactor
 refactored
 refactors
 registerExtension
+rescan
 resettable
 richeland
 RSS

docs/changelog.md

@@ -23,6 +23,8 @@ See the [Contributing Guide](contributing.md) for details.
   consisting only of `{` and some text with many trailing spaces (#1656).
 * Fix quadratic-time regex backtracking in hash headers when a long run of `#`
   characters is followed by more text on the same line.
+* Fix quadratic time when a document contains many unclosed HTML comments (`<!--`
+  with no matching `-->`).
 * Fix regression in `BacktickInlineProcessor` where text could be lost when opening
   and closing backticks did not match (#1668).
 * Escape the language name in the `codehilite` extension's non-Pygments output to

markdown/htmlparser.py

@@ -158,6 +158,8 @@ def reset(self):
         self._cache: list[str] = []
         self.cleandoc: list[str] = []
         self.lineno_start_cache = [0]
+        # The `rawdata` and position of the last failed search for a comment close.
+        self._no_commentclose: tuple[str | None, int] = (None, 0)
 
         super().reset()
 
@@ -359,8 +361,15 @@ def parse_pi(self, i: int) -> int:
     def parse_comment(self, i, report=True):
         rawdata = self.rawdata
         assert rawdata.startswith('<!--', i), 'unexpected call to parse_comment()'
-        match = commentclose.search(rawdata, i+4)
+        # If no close was found after an earlier position in the same `rawdata`, there
+        # is none after this one either. Don't rescan the rest of the document again.
+        searched, pos = self._no_commentclose
+        if searched is rawdata and i + 4 >= pos:
+            match = None
+        else:
+            match = commentclose.search(rawdata, i+4)
         if not match:
+            self._no_commentclose = (rawdata, i + 4)
             self.handle_data('<')
             return i + 1
         if report:

tests/test_syntax/blocks/test_html_blocks.py

@@ -1860,6 +1860,13 @@ def test_multiple_unclosed_comments_no_hang(self):
             '<p>&lt;!-- and &lt;!--</p>'
         )
 
+    def test_many_unclosed_comments(self):
+        """Test that each unclosed comment does not rescan the rest of the document."""
+        self.assertMarkdownRenders(
+            '<!--' * 50000,
+            '<p>' + '&lt;!--' * 50000 + '</p>'
+        )
+
     def test_no_hang_issue_1586(self):
         """Test no hang condition for issue #1586."""