markdown: stricter detection of embedded markup

instead of only checking only for the existence of `<`, check for text
directly behind the opening bracket and a closing bracket.

This prevents added paragraphs/new lines with just brackets or "arrows"
(`<-`) in lines.

fixes #1761

src/moin/converters/_tests/test_markdown_in.py

@@ -182,6 +182,27 @@ def test_wikilinks(self, input, output):
     def test_admonition(self, input, output):
         self.do(input, output)
 
+    data = [
+        (
+            "one < two",
+            '<p>one &lt; two</p>',
+        ),
+        (
+            '[[one]] < two',
+            '<p><a xlink:href="wiki.local:one">one</a> &lt; two</p>',
+        ),
+        (
+            'pre <strong>bold</strong> post',
+            '<div><p>pre <strong>bold</strong> post</p></div>',
+        ),
+    ]
+
+    @pytest.mark.parametrize("input,output", data)
+    def test_embedded_markup(self, input, output):
+        """Test embedded markup in markdown"""
+        self.do(input, output)
+
+
     def serialize_strip(self, elem, **options):
         result = serialize(elem, namespaces=self.namespaces, **options)
         return self.output_re.sub("", result)

src/moin/converters/markdown_in.py

@@ -531,7 +531,8 @@ def convert_embedded_markup(self, node):
         """
         for idx, child in enumerate(node):
             if isinstance(child, str):
-                if "<" in child:
+                # search for HTML tags
+                if re.search("<[^ ].*?>", child):
                     node[idx] = self.embedded_markup(child)  # child is immutable string, so must do node[idx]
             else:
                 # do not convert markup within a <pre> tag