Merge pull request #713 from apoorvdarshan/fix/header-ids-duplicate-661

Fix header-ids extra generating duplicate ids (#661)

Author: nicholasserra

CHANGES.md

@@ -9,6 +9,7 @@
 - [pull #701] Allow boolean attribute syntax in `markdown-in-html` extra
 - [pull #704] Fix XSS from smuggling spans into image attributes (#702, #703)
 - [pull #710] Add emoji support (#709)
+- [pull #713] Fix `header-ids` extra generating duplicate ids when a suffixed id collides with another header (#661)
 
 
 ## python-markdown2 2.5.5

lib/markdown2.py

@@ -413,6 +413,8 @@ def _setup_extras(self):
         if "header-ids" in self.extras:
             if not hasattr(self, '_count_from_header_id') or self.extras['header-ids'].get('reset-count', False):
                 self._count_from_header_id = defaultdict(int)
+            if not hasattr(self, '_header_ids_seen') or self.extras['header-ids'].get('reset-count', False):
+                self._header_ids_seen = set()
         if "metadata" in self.extras:
             self.metadata: dict[str, Any] = {}
 
@@ -1583,9 +1585,17 @@ def header_id_from_text(self,
         if prefix and isinstance(prefix, str):
             header_id = prefix + '-' + header_id
 
-        self._count_from_header_id[header_id] += 1
-        if 0 == len(header_id) or self._count_from_header_id[header_id] > 1:
-            header_id += '-%s' % self._count_from_header_id[header_id]
+        base_id = header_id
+        self._count_from_header_id[base_id] += 1
+        if 0 == len(base_id) or self._count_from_header_id[base_id] > 1:
+            header_id = '%s-%s' % (base_id, self._count_from_header_id[base_id])
+        # A suffixed id may still collide with a differently-named header
+        # (e.g. "# Chapter" twice yields "chapter-2", which clashes with
+        # "# Chapter 2"). Keep bumping until the id is genuinely unique.
+        while header_id in self._header_ids_seen:
+            self._count_from_header_id[base_id] += 1
+            header_id = '%s-%s' % (base_id, self._count_from_header_id[base_id])
+        self._header_ids_seen.add(header_id)
 
         return header_id
 

test/tm-cases/header_ids_duplicate.html

@@ -0,0 +1,11 @@
+<h1 id="chapter">Chapter</h1>
+
+<p>Test</p>
+
+<h1 id="chapter-2">Chapter</h1>
+
+<p>Test</p>
+
+<h1 id="chapter-2-2">Chapter 2</h1>
+
+<p>Test</p>

test/tm-cases/header_ids_duplicate.opts

@@ -0,0 +1 @@
+{"extras": ["header-ids"]}

test/tm-cases/header_ids_duplicate.tags

@@ -0,0 +1 @@
+extra header-ids issue661

test/tm-cases/header_ids_duplicate.text

@@ -0,0 +1,11 @@
+# Chapter
+
+Test
+
+# Chapter
+
+Test
+
+# Chapter 2
+
+Test