fix(jruby): XHTML formatting of non-container element closing tags

The JRuby code for SaveContextVisitor is brittle and clearly buggy (since this bug exists for fragment nodes but not fragments or document nodes?), but this seems like an easy fix. Fixes #2355
sparklemotion · Nov 6, 2021 · f7d847f · f7d847f
1 parent 3800769
commit f7d847f
Show file tree

Hide file tree

Showing 3 changed files with 9 additions and 0 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -20,6 +20,7 @@ Nokogiri follows [Semantic Versioning](https://semver.org/), please see the [REA
 * [CRuby] Fix memory leak in XPath custom handlers where string arguments were not being cleaned up. [[#2345](https://github.com/sparklemotion/nokogiri/issues/2345)]
 * [CRuby] Fix memory leak in `Reader#base_uri` where the string returned by libxml2 was not freed. [[#2347](https://github.com/sparklemotion/nokogiri/issues/2347)]
 * [JRuby] Deleting a `Namespace` from a `NodeSet` no longer modifies the `href` to be the default namespace URL.
+* [JRuby] Fix XHTML formatting of closing tags for non-container elements. [[#2355](https://github.com/sparklemotion/nokogiri/issues/2355)]
 
 
 ### Improved

diff --git a/ext/java/nokogiri/internals/SaveContextVisitor.java b/ext/java/nokogiri/internals/SaveContextVisitor.java
@@ -673,6 +673,8 @@ public int compare(Attr attr0, Attr attr1) {
       if (!isEmpty(name) && noEmpty) {
         buffer.append("</").append(name).append('>');
       }
+    } else if (asXhtml && !isEmpty(name)) {
+      buffer.append("</").append(name).append('>');
     }
     if (needBreakInClosing(element)) {
       if (!containsText(element)) { indentation.pop(); }

diff --git a/test/html4/test_node.rb b/test/html4/test_node.rb
@@ -203,6 +203,12 @@ def test_GH_1042
         # are not set. the fix for GH1042 ensures a proper working clone.
         trs.inspect # assert_nothing_raised
       end
+
+      def test_fragment_node_to_xhtml # see #2355
+        html = "<h1><a></a><br>First H1</h1>"
+        fragment = Nokogiri::HTML4::DocumentFragment.parse(html)
+        assert_equal("<h1><a></a><br />First H1</h1>", fragment.at_css("h1").to_xhtml)
+      end
     end
   end
 end