Fix sanitizer parsing link text as HTML when stripping unsupported links (#22558) (#23491)

lib/sanitize_ext/sanitize_config.rb

@@ -49,7 +49,7 @@ class Sanitize
         end
       end
 
-      current_node.replace(current_node.text) unless LINK_PROTOCOLS.include?(scheme)
+      current_node.replace(Nokogiri::XML::Text.new(current_node.text, current_node.document)) unless LINK_PROTOCOLS.include?(scheme)
     end
 
     UNSUPPORTED_ELEMENTS_TRANSFORMER = lambda do |env|

spec/lib/sanitize_config_spec.rb

@@ -38,6 +38,10 @@ describe Sanitize::Config do
       expect(Sanitize.fragment('<a href="foo://bar">Test</a>', subject)).to eq 'Test'
     end
 
+    it 'does not re-interpret HTML when removing unsupported links' do
+      expect(Sanitize.fragment('<a href="foo://bar">Test&lt;a href="https://example.com"&gt;test&lt;/a&gt;</a>', subject)).to eq 'Test&lt;a href="https://example.com"&gt;test&lt;/a&gt;'
+    end
+
     it 'keeps a with href' do
       expect(Sanitize.fragment('<a href="http://example.com">Test</a>', subject)).to eq '<a href="http://example.com" rel="nofollow noopener noreferrer" target="_blank">Test</a>'
     end