Pre parse html with nokogiri to fix unclosed tags

html2confluence.gemspec

@@ -1,7 +1,7 @@
 Gem::Specification.new do |s|
   s.platform    = Gem::Platform::RUBY
   s.name        = 'html2confluence'
-  s.version     = "1.3.8"
+  s.version     = "1.3.9"
   s.summary     = 'Converter from HTML to Confluence Wiki Markup'
   s.description = 'Provides an SGML parser to convert HTML into the Wiki Markup format'
 
@@ -13,4 +13,6 @@ Gem::Specification.new do |s|
 
   s.require_path = 'lib'
   s.files        = Dir.glob("{lib,spec}/**/*") + %w(example.rb README.mdown)
+
+  s.add_dependency "nokogiri"
 end

lib/html2confluence.rb

@@ -1,5 +1,7 @@
 require 'rexml/document'
 
+require 'nokogiri' # For validating html from our editor
+
 # A class to convert HTML to textile. Based on the python parser
 # found at http://aftnn.org/content/code/html2textile/
 #
@@ -406,6 +408,14 @@ def preprocess(data)
 
     # Fix unclosed <img>
     data.gsub!(/(<img[^>]+)(?<!\/)>/, '\1 />')
+
+    # Parse with nokogiri to ensure not tags are left unclosed
+    # Ensure a parsing error from Nokogiri can't stop processing to get better error from REXML
+    begin
+      validated_data = Nokogiri::HTML::fragment(data).to_xml
+      data = validated_data
+    rescue Exception => e
+    end
 
     data
   end