Fix unescaping of Markdown titles

Currently this is a regression in unreleased code. Extensions such as 'inlinehilite' and 'smarty' were pasting `wzxhzdk:1` due to their use of htmlStash

mkdocs/structure/pages.py

@@ -4,13 +4,11 @@
 import logging
 import posixpath
 import warnings
-from typing import TYPE_CHECKING, Any, Callable, Iterator, MutableMapping
+from typing import TYPE_CHECKING, Any, Callable, Iterator, MutableMapping, Sequence
 from urllib.parse import unquote as urlunquote
 from urllib.parse import urljoin, urlsplit, urlunsplit
 
 import markdown
-import markdown.extensions
-import markdown.postprocessors
 import markdown.treeprocessors
 from markdown.util import AMP_SUBSTITUTE
 
@@ -22,16 +20,12 @@
 if TYPE_CHECKING:
     from xml.etree import ElementTree as etree
 
+    import markdown.postprocessors
+
     from mkdocs.config.defaults import MkDocsConfig
     from mkdocs.structure.files import File, Files
     from mkdocs.structure.toc import TableOfContents
 
-_unescape: Callable[[str], str]
-try:
-    _unescape = markdown.treeprocessors.UnescapeTreeprocessor().unescape  # type: ignore
-except AttributeError:
-    _unescape = markdown.postprocessors.UnescapePostprocessor().run
-
 
 log = logging.getLogger(__name__)
 
@@ -437,6 +431,7 @@ def _register(self, md: markdown.Markdown) -> None:
 
 class _ExtractTitleTreeprocessor(markdown.treeprocessors.Treeprocessor):
     title: str | None = None
+    postprocessors: Sequence[markdown.postprocessors.Postprocessor] = ()
 
     def run(self, root: etree.Element) -> etree.Element:
         for el in root:
@@ -446,13 +441,18 @@ def run(self, root: etree.Element) -> etree.Element:
                     el = copy.copy(el)
                     del el[-1]
                 # Extract the text only, recursively.
-                self.title = _unescape(''.join(el.itertext()))
+                title = ''.join(el.itertext())
+                # Unescape per Markdown implementation details.
+                for pp in self.postprocessors:
+                    title = pp.run(title)
+                self.title = title
             break
         return root
 
     def _register(self, md: markdown.Markdown) -> None:
+        self.postprocessors = tuple(md.postprocessors)
         md.treeprocessors.register(
             self,
             "mkdocs_extract_title",
-            priority=1,  # Close to the end.
+            priority=-1,  # After the end.
         )

mkdocs/tests/structure/page_tests.py

@@ -352,7 +352,7 @@ def test_page_title_from_markdown_stripped_anchorlinks(self, docs_dir):
 
     _FORMATTING_CONTENT = dedent(
         '''
-        # Hello *beautiful* `world`
+        # \\*Hello --- *beautiful* `world`
 
         Hi.
         '''
@@ -361,11 +361,12 @@ def test_page_title_from_markdown_stripped_anchorlinks(self, docs_dir):
     @tempdir(files={'testing_formatting.md': _FORMATTING_CONTENT})
     def test_page_title_from_markdown_strip_formatting(self, docs_dir):
         cfg = load_config()
+        cfg.markdown_extensions.append('smarty')
         fl = File('testing_formatting.md', docs_dir, docs_dir, use_directory_urls=True)
         pg = Page(None, fl, cfg)
         pg.read_source(cfg)
         pg.render(cfg, fl)
-        self.assertEqual(pg.title, 'Hello beautiful world')
+        self.assertEqual(pg.title, '*Hello — beautiful world')
 
     _ATTRLIST_CONTENT = dedent(
         '''