Strip all HTML when getting the title from the first H1 tag (#3564)

Not stripping it was a bug, and also inconsistent with how ToC titles are extracted.

mkdocs/structure/pages.py

@@ -10,6 +10,7 @@ from urllib.parse import unquote as urlunquote
 from urllib.parse import urljoin, urlsplit, urlunsplit
 
 import markdown
+import markdown.extensions.toc
 import markdown.htmlparser  # type: ignore
 import markdown.postprocessors
 import markdown.treeprocessors
@@ -549,7 +550,7 @@ class _HTMLHandler(markdown.htmlparser.htmlparser.HTMLParser):  # type: ignore[n
 
 class _ExtractTitleTreeprocessor(markdown.treeprocessors.Treeprocessor):
     title: str | None = None
-    postprocessors: Sequence[markdown.postprocessors.Postprocessor] = ()
+    md: markdown.Markdown
 
     def run(self, root: etree.Element) -> etree.Element:
         for el in root:
@@ -561,14 +562,15 @@ class _ExtractTitleTreeprocessor(markdown.treeprocessors.Treeprocessor):
                 # Extract the text only, recursively.
                 title = ''.join(el.itertext())
                 # Unescape per Markdown implementation details.
-                for pp in self.postprocessors:
+                title = markdown.extensions.toc.stashedHTML2text(
-                    title = pp.run(title)
+                    title, self.md, strip_entities=False
-                self.title = title
+                )
+                self.title = title.strip()
             break
         return root
 
     def _register(self, md: markdown.Markdown) -> None:
-        self.postprocessors = tuple(md.postprocessors)
+        self.md = md
         md.treeprocessors.register(self, "mkdocs_extract_title", priority=-1)  # After the end.
 
 

mkdocs/structure/toc.py

@@ -33,7 +33,7 @@ class AnchorLink:
         self.children = []
 
     title: str
-    """The text of the item."""
+    """The text of the item, as HTML."""
 
     @property
     def url(self) -> str:

mkdocs/tests/structure/page_tests.py

@@ -6,9 +6,11 @@ import textwrap
 import unittest
 from unittest import mock
 
+import markdown
+
 from mkdocs.config.defaults import MkDocsConfig
 from mkdocs.structure.files import File, Files
-from mkdocs.structure.pages import Page, _RelativePathTreeprocessor
+from mkdocs.structure.pages import Page, _ExtractTitleTreeprocessor, _RelativePathTreeprocessor
 from mkdocs.tests.base import dedent, tempdir
 
 DOCS_DIR = os.path.join(
@@ -315,9 +317,16 @@ class PageTests(unittest.TestCase):
         self.assertEqual(pg.parent, None)
         self.assertEqual(pg.previous_page, None)
         self.assertEqual(pg.title, 'Welcome to MkDocs')
-        pg.render(cfg, fl)
+        pg.render(cfg, Files([fl]))
         self.assertEqual(pg.title, 'Welcome to MkDocs')
 
+    def _test_extract_title(self, content, expected, extensions={}):
+        md = markdown.Markdown(extensions=list(extensions.keys()), extension_configs=extensions)
+        extract_title_ext = _ExtractTitleTreeprocessor()
+        extract_title_ext._register(md)
+        md.convert(content)
+        self.assertEqual(extract_title_ext.title, expected)
+
     _SETEXT_CONTENT = dedent(
         '''
         Welcome to MkDocs Setext
@@ -327,46 +336,37 @@ class PageTests(unittest.TestCase):
         '''
     )
 
-    @tempdir(files={'testing_setext_title.md': _SETEXT_CONTENT})
+    def test_page_title_from_setext_markdown(self):
-    def test_page_title_from_setext_markdown(self, docs_dir):
+        self._test_extract_title(
-        cfg = load_config()
+            self._SETEXT_CONTENT,
-        fl = File('testing_setext_title.md', docs_dir, docs_dir, use_directory_urls=True)
+            expected='Welcome to MkDocs Setext',
-        pg = Page(None, fl, cfg)
+        )
-        self.assertIsNone(pg.title)
-        pg.read_source(cfg)
-        self.assertEqual(pg.title, 'Testing setext title')
-        pg.render(cfg, fl)
-        self.assertEqual(pg.title, 'Welcome to MkDocs Setext')
 
-    @tempdir(files={'testing_setext_title.md': _SETEXT_CONTENT})
+    def test_page_title_from_markdown_stripped_anchorlinks(self):
-    def test_page_title_from_markdown_stripped_anchorlinks(self, docs_dir):
+        self._test_extract_title(
-        cfg = MkDocsConfig()
+            self._SETEXT_CONTENT,
-        cfg.site_name = 'example'
+            extensions={'toc': {'permalink': '&'}},
-        cfg.markdown_extensions = {'toc': {'permalink': '&'}}
+            expected='Welcome to MkDocs Setext',
-        self.assertEqual(cfg.validate(), ([], []))
+        )
-        fl = File('testing_setext_title.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, 'Welcome to MkDocs Setext')
 
-    _FORMATTING_CONTENT = dedent(
+    def test_page_title_from_markdown_strip_formatting(self):
-        '''
+        self._test_extract_title(
-        # \\*Hello --- *beautiful* `world`
+            '''# \\*Hello --- *beautiful* `wor<dl>`''',
+            extensions={'smarty': {}},
+            expected='*Hello &mdash; beautiful wor&lt;dl&gt;',
+        )
 
-        Hi.
+    def test_page_title_from_markdown_strip_raw_html(self):
-        '''
+        self._test_extract_title(
-    )
+            '''# Hello <b>world</b>''',
+            expected='Hello world',
+        )
 
-    @tempdir(files={'testing_formatting.md': _FORMATTING_CONTENT})
+    def test_page_title_from_markdown_strip_image(self):
-    def test_page_title_from_markdown_strip_formatting(self, docs_dir):
+        self._test_extract_title(
-        cfg = load_config()
+            '''# Hi ![😄](hah.png)''',
-        cfg.markdown_extensions.append('smarty')
+            expected='Hi',  # TODO: Should the alt text of the image be extracted?
-        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 &mdash; beautiful world')
 
     _ATTRLIST_CONTENT = dedent(
         '''
@@ -376,24 +376,18 @@ class PageTests(unittest.TestCase):
         '''
     )
 
-    @tempdir(files={'testing_attr_list.md': _ATTRLIST_CONTENT})
+    def test_page_title_from_markdown_stripped_attr_list(self):
-    def test_page_title_from_markdown_stripped_attr_list(self, docs_dir):
+        self._test_extract_title(
-        cfg = load_config()
+            self._ATTRLIST_CONTENT,
-        cfg.markdown_extensions.append('attr_list')
+            extensions={'attr_list': {}},
-        fl = File('testing_attr_list.md', docs_dir, docs_dir, use_directory_urls=True)
+            expected='Welcome to MkDocs Attr',
-        pg = Page(None, fl, cfg)
+        )
-        pg.read_source(cfg)
-        pg.render(cfg, fl)
-        self.assertEqual(pg.title, 'Welcome to MkDocs Attr')
 
-    @tempdir(files={'testing_attr_list.md': _ATTRLIST_CONTENT})
+    def test_page_title_from_markdown_preserved_attr_list(self):
-    def test_page_title_from_markdown_preserved_attr_list(self, docs_dir):
+        self._test_extract_title(
-        cfg = load_config()
+            self._ATTRLIST_CONTENT,
-        fl = File('testing_attr_list.md', docs_dir, docs_dir, use_directory_urls=True)
+            expected='Welcome to MkDocs Attr { #welcome }',
-        pg = Page(None, fl, cfg)
+        )
-        pg.read_source(cfg)
-        pg.render(cfg, fl)
-        self.assertEqual(pg.title, 'Welcome to MkDocs Attr { #welcome }')
 
     def test_page_title_from_meta(self):
         cfg = load_config(docs_dir=DOCS_DIR)
@@ -418,7 +412,7 @@ class PageTests(unittest.TestCase):
         self.assertEqual(pg.previous_page, None)
         self.assertEqual(pg.title, 'A Page Title')
         self.assertEqual(pg.toc, [])
-        pg.render(cfg, fl)
+        pg.render(cfg, Files([fl]))
         self.assertEqual(pg.title, 'A Page Title')
 
     def test_page_title_from_filename(self):
@@ -443,7 +437,7 @@ class PageTests(unittest.TestCase):
         self.assertEqual(pg.parent, None)
         self.assertEqual(pg.previous_page, None)
         self.assertEqual(pg.title, 'Page title')
-        pg.render(cfg, fl)
+        pg.render(cfg, Files([fl]))
         self.assertEqual(pg.title, 'Page title')
 
     def test_page_title_from_capitalized_filename(self):
@@ -704,7 +698,7 @@ class PageTests(unittest.TestCase):
         pg.read_source(cfg)
         self.assertEqual(pg.content, None)
         self.assertEqual(pg.toc, [])
-        pg.render(cfg, [fl])
+        pg.render(cfg, Files([fl]))
         self.assertTrue(
             pg.content.startswith('<h1 id="welcome-to-mkdocs">Welcome to MkDocs</h1>\n')
         )

pyproject.toml

@@ -36,7 +36,7 @@ dependencies = [
     "click >=7.0",
     "Jinja2 >=2.11.1",
     "markupsafe >=2.0.1",
-    "Markdown >=3.3.6",
+    "Markdown >=3.4.1",
     "PyYAML >=5.1",
     "watchdog >=2.0",
     "ghp-import >=1.0",
@@ -57,7 +57,7 @@ min-versions = [
     "click ==7.0",
     "Jinja2 ==2.11.1",
     "markupsafe ==2.0.1",
-    "Markdown ==3.3.6",
+    "Markdown ==3.4.1",
     "PyYAML ==5.1",
     "watchdog ==2.0",
     "ghp-import ==1.0",