Add support for intentionally empty tags

lib/htmldiff/diff_builder.rb

@@ -16,10 +16,15 @@ def initialize(old_version, new_version, options = {})
 
     def default_options
       {
-        block_tag_classes: []
+        block_tag_classes: [],
+        compare_tag_attributes: false
       }
     end
 
+    def compare_tag_attributes?
+      @options[:compare_tag_attributes]
+    end
+
     def build
       perform_operations
       content.join
@@ -54,7 +59,8 @@ def replace(operation)
       # added e.g. <p> becomes <p style="margin: 2px"> due to an editor button
       # press. For this, we just show the new version, otherwise it gets messy
       # trying to find the closing tag.
-      if operation.same_tag?
+
+      if operation.same_tag?(compare_tag_attributes?)
         equal(operation)
       else
         delete(operation, 'diffmod')
@@ -99,10 +105,12 @@ def insert_tag(tagname, cssclass, words)
       loop do
         break if words.empty?
 
-        if words.first.standalone_tag?
-          tag_words = words.extract_consecutive_words! do |word|
-            word.standalone_tag?
-          end
+        # Handle empty tags as single blocks
+        if words.first.closed_empty_tag?
+          tag_words = words.extract_consecutive_words! { |word| word.closed_empty_tag? }
+          @content << wrap_text_in_diff_tag(tag_words.join, tagname, cssclass)
+        elsif words.first.standalone_tag?
+          tag_words = words.extract_consecutive_words! { |word| word.standalone_tag? }
           @content << wrap_text_in_diff_tag(tag_words.join, tagname, cssclass)
         elsif words.first.iframe_tag?
           tag_words = words.extract_consecutive_words! { |word| word.iframe_tag? }
@@ -124,7 +132,7 @@ def insert_tag(tagname, cssclass, words)
             wrapped = true
           end
           @content += words.extract_consecutive_words! do |word|
-            word.tag? && !word.standalone_tag? && !word.iframe_tag?
+            word.tag? && !word.standalone_tag? && !word.iframe_tag? && !word.closed_empty_tag?
           end
         else
           non_tags = words.extract_consecutive_words! do |word|

lib/htmldiff/list_of_words.rb

@@ -13,6 +13,7 @@ def initialize(string, options = {})
         @words = string
       else
         convert_html_to_list_of_words string.chars
+        group_empty_tags!
       end
     end
 
@@ -72,6 +73,55 @@ def contains_unclosed_tag?
 
     private
 
+    def group_empty_tags!
+      return if @words.empty?
+      new_words = []
+      i = 0
+
+      while i < @words.length
+        current_word = @words[i]
+
+        # Check if this is an opening tag
+        if (tag_match = current_word.to_s.match(/^<([^\s>\/]+)[^>]*>$/i))
+          tag_name = tag_match[1]
+
+          # Look ahead to see if the very next word (after any whitespace) is the closing tag
+          # next_non_whitespace_index = find_next_non_whitespace_word(i + 1)
+          next_index = i + 1
+          # If the very next word is the closing tag, group the empty tag pair
+          if @words[next_index]&.to_s&.match?(/^<\/#{Regexp.escape(tag_name)}>$/i)
+            word_group = []
+            (i..next_index).each do |idx|
+              word_group << @words[idx]
+            end
+
+            new_words << Word.new(word_group.map(&:to_s).join)
+            i = next_index + 1
+          else
+            # Otherwise, add as individual word
+            new_words << current_word
+            i += 1
+          end
+        else
+          # Not an opening tag - keep individual word
+          new_words << current_word
+          i += 1
+        end
+      end
+
+      @words = new_words
+    end
+
+    def find_next_non_whitespace_word(start_index)
+      i = start_index
+      while i < @words.length
+        word_str = @words[i].to_s.strip
+        return i unless word_str.empty?
+        i += 1
+      end
+      nil
+    end
+
     def convert_html_to_list_of_words(character_array)
       @mode = :char
       @current_word = Word.new

lib/htmldiff/operation.rb

@@ -14,17 +14,25 @@ class Operation
     # @!method old_words
     # @!method new_words
 
-    # Ignores any attributes and tells us if the tag is the same e.g. <p> and
-    # <p style="margin: 2px;"> are the same.
-    def same_tag?
-      pattern = /<([^>\s]+)[\s>].*/
-      first_tagname = pattern.match(old_text) # nil means they are not tags
-      first_tagname = first_tagname[1] if first_tagname
+    # Unless `compare_attributes` is true, Ignores any attributes and tells us
+    # if the tag is the same e.g. <p> and <p style="margin: 2px;"> are the same.
+    def same_tag?(compare_attributes = false)
+      pattern = /<(?<name>[^>\s]+)\s(?<attrs>.*)[\s>].*/
+      first_tag = pattern.match(old_text) # nil means they are not tags
+      first_tagname = first_tag[:name] if first_tag
 
-      second_tagname = pattern.match(new_text)
-      second_tagname = second_tagname[1] if second_tagname
+      second_tag = pattern.match(new_text)
+      second_tagname = second_tag[:name] if second_tag
 
-      first_tagname && (first_tagname == second_tagname)
+      return false unless first_tag && second_tag
+
+      if compare_attributes
+        first_attrs = first_tag[:attrs]
+        second_attrs = first_tag[:attrs]
+        return false if first_attrs && (first_attrs == second_attrs)
+      end
+
+      first_tagname == second_tagname
     end
 
     def old_text

lib/htmldiff/word.rb

@@ -26,6 +26,10 @@ def iframe_tag?
       (@word[0..7].downcase =~ %r{^<\/?iframe ?})
     end
 
+    def closed_empty_tag?
+      @word.match?(/^<([^\/\s>]+)(?:\s[^>]*)?>(?:\s*)<\/\1>$/)
+    end
+
     def tag?
       opening_tag? || closing_tag? || standalone_tag?
     end