Merge pull request #576 from Crozzers/fix-575

Fix `html`, `head` and `body` tags being wrapped in `<p>` tags (#575)
trentm · Apr 22, 2024 · b2a512d · b2a512d
2 parents ac7fd19 + 0e51afb
commit b2a512d
Show file tree

Hide file tree

Showing 4 changed files with 49 additions and 2 deletions.
diff --git a/CHANGES.md b/CHANGES.md
@@ -9,6 +9,7 @@
 - [pull #570] Fix syntax warnings in test suite
 - [pull #572] Process inline tags as HTML blocks when they span multiple lines (#571)
 - [pull #573] Add new LaTeX Extra
+- [pull #576] Fix `html`, `head` and `body` tags being wrapped in `<p>` tags (#575)
 
 
 ## python-markdown2 2.4.13

diff --git a/lib/markdown2.py b/lib/markdown2.py
@@ -814,7 +814,7 @@ def _detab(self, text):
     # _block_tags_b.  This way html5 tags are easy to keep track of.
     _html5tags = '|article|aside|header|hgroup|footer|nav|section|figure|figcaption'
 
-    _block_tags_a = 'p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math|ins|del|style'
+    _block_tags_a = 'p|div|h[1-6]|blockquote|pre|table|dl|ol|ul|script|noscript|form|fieldset|iframe|math|ins|del|style|html|head|body'
     _block_tags_a += _html5tags
 
     _strict_tag_block_re = re.compile(r"""
@@ -2773,7 +2773,7 @@ class Latex(Extra):
 
     _single_dollar_re = re.compile(r'(?<!\$)\$(?!\$)(.*?)\$')
     _double_dollar_re = re.compile(r'\$\$(.*?)\$\$', re.DOTALL)
-    
+
     # Ways to escape
     _pre_code_block_re = re.compile(r"<pre>(.*?)</pre>", re.DOTALL) # Wraped in <pre>
     _triple_re = re.compile(r'```(.*?)```', re.DOTALL) # Wrapped in a code block ```

diff --git a/test/tm-cases/block_tags.html b/test/tm-cases/block_tags.html
@@ -0,0 +1,23 @@
+<html>
+
+<!-- Comment please ignore -->
+
+<!-- Multi
+line Comment
+please ignore -->
+
+
+<body>
+content here
+
+ <img src="some_img.jpg" alt="there is supposed to be an image here" width="500" height="600">
+
+Now some bullets:
+
+  * one
+  * two
+
+</body>
+
+
+</html>
diff --git a/test/tm-cases/block_tags.text b/test/tm-cases/block_tags.text
@@ -0,0 +1,23 @@
+<html>
+
+<!-- Comment please ignore -->
+
+<!-- Multi
+line Comment
+please ignore -->
+
+
+<body>
+content here
+
+ <img src="some_img.jpg" alt="there is supposed to be an image here" width="500" height="600">
+
+Now some bullets:
+
+  * one
+  * two
+
+</body>
+
+
+</html>