Keep wiki links out of existing links and HTML attributes

CHANGES.md

@@ -2,6 +2,7 @@
 
 ## python-markdown2 2.5.6 (not yet released)
 
+- Fix `wiki-links` generating nested anchors and replacing syntax inside image descriptions and HTML attributes.
 - [pull #762] Speed up documents with many links by skipping the backslash-escape scan for text without a backslash
 - [pull #746] Preserve query strings and fragments in `markdown-file-links` conversions, including reference-link targets.
 - [pull #739] Fix `header-ids` extra putting sanitized HTML into header ids in safe mode, which allowed attribute injection (#649)

lib/markdown2.py

@@ -4482,6 +4482,8 @@ class WikiLinks(Extra):
     '''
     Wiki-style links. `[[Page Name]]` becomes a link to another page. This is
     modelled on the wikilinks extension shipped with Python-Markdown.
+    Wiki syntax inside existing links, image descriptions, or HTML attributes
+    is left as literal text.
 
     The generated URL and CSS class can be tuned with the following options:
 
@@ -4500,12 +4502,24 @@ class WikiLinks(Extra):
     yields ``<a href="/Foo_Bar.html" class="localLink">Foo Bar</a>``.
     '''
     name = 'wiki-links'
-    order = (Stage.LINKS,), ()
+    order = (), (Stage.LINKS,)
 
     _wiki_link_re = re.compile(r'\[\[([\w0-9_ -]+)\]\]', re.UNICODE)
 
     def run(self, text: str) -> str:
-        return self._wiki_link_re.sub(self.sub, text)
+        # Resolve Markdown links first, then only link text outside existing
+        # anchors and HTML markup (including image attributes).
+        tokens = self.md._sorta_html_tokenize_re.split(text)
+        in_anchor = False
+        for index, token in enumerate(tokens):
+            if index % 2:
+                if re.match(r'<a(?:\s|>)', token, re.I):
+                    in_anchor = True
+                elif re.match(r'</a\s*>', token, re.I):
+                    in_anchor = False
+            elif not in_anchor:
+                tokens[index] = self._wiki_link_re.sub(self.sub, token)
+        return ''.join(tokens)
 
     def build_url(self, label: str) -> str:
         base_url = self.options.get('base_url', '/')

test/test_markdown2.py

@@ -220,6 +220,33 @@ class DirectTestCase(_MarkdownTestCase):
     Python-markdown (markdown.py).
     """
 
+    def test_wiki_links_in_links_and_images(self):
+        cases = [
+            ('[go [[Home]]](https://example.com)',
+             '<p><a href="https://example.com">go [[Home]]</a></p>\n'),
+            ('![go [[Home]]](img.png)',
+             '<p><img src="img.png" alt="go [[Home]]" /></p>\n'),
+            ('[go [[Home]]][target]\n\n[target]: https://example.com',
+             '<p><a href="https://example.com">go [[Home]]</a></p>\n'),
+            ('<a href="/old">[[Home]]</a>',
+             '<p><a href="/old">[[Home]]</a></p>\n'),
+            ('<A HREF="/old"><em>[[Home]]</em></A> [[Page]]',
+             '<p><A HREF="/old"><em>[[Home]]</em></A> <a href="/Page/" class="wikilink">Page</a></p>\n'),
+            ('![go [[Home]]][img]\n\n[img]: pic.png',
+             '<p><img src="pic.png" alt="go [[Home]]" /></p>\n'),
+            ('[![go [[Home]]](img.png)](/go)',
+             '<p><a href="/go"><img src="img.png" alt="go [[Home]]" /></a></p>\n'),
+            (r'\[[Home]] `[[Home]]` [[Page]]',
+             '<p>[[Home]] <code>[[Home]]</code> <a href="/Page/" class="wikilink">Page</a></p>\n'),
+            ('<span title="[[Home]]">[[Page]]</span>',
+             '<p><span title="[[Home]]"><a href="/Page/" class="wikilink">Page</a></span></p>\n'),
+            ('[go](https://example.com) [[Home]]',
+             '<p><a href="https://example.com">go</a> <a href="/Home/" class="wikilink">Home</a></p>\n'),
+        ]
+        for source, expected in cases:
+            with self.subTest(source=source):
+                self.assertEqual(markdown2.markdown(source, extras=['wiki-links']), expected)
+
     def test_header_ids_from_file_vars(self):
         md = markdown2.Markdown(use_file_vars=True)
         modeline = "<!-- -*- markdown-extras: header-ids{} -*- -->\n"