Add support for intentionally empty tags

lib/htmldiff/diff_builder.rb

@@ -54,6 +54,7 @@ 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.
+      # our-embed tags are the exception and will be replaced in full.
       if operation.same_tag?
         equal(operation)
       else
@@ -99,10 +100,12 @@ 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 our-embeds and writing blank spans as single blocks
+        if words.first.closed_embed_or_blank_tag?
+          tag_words = words.extract_consecutive_words! { |word| word.closed_embed_or_blank_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 +127,7 @@ 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_embed_or_blank_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_embed_or_blank_tags!
       end
     end
 
@@ -72,6 +73,43 @@ def contains_unclosed_tag?
 
     private
 
+    # Group our-embed tags and Writing Blank spans, which are
+    # intentionally left blank, into single words
+    def group_embed_or_blank_tags!
+      return if @words.empty?
+      new_words = []
+      i = 0
+
+      while i < @words.length
+        current_word = @words[i]
+
+        if current_word.embed_or_blank_opening_tag?
+          word_group = [current_word]
+          tag_name = current_word.to_s.match(/^<(span|our-embed)/)[1]
+          i += 1
+
+          # Collect words until the appropriate closing tag is reached
+          while i < @words.length
+            word = @words[i]
+            word_group << word
+
+            if word.to_s.match?(/^<\/#{tag_name}>$/)
+              i += 1
+              break
+            end
+
+            i += 1
+          end
+          # Create a single word from the entire element
+          new_words << Word.new(word_group.map(&:to_s).join)
+        else
+          new_words << current_word
+          i += 1
+        end
+      end
+      @words = new_words
+    end
+
     def convert_html_to_list_of_words(character_array)
       @mode = :char
       @current_word = Word.new

lib/htmldiff/operation.rb

@@ -16,7 +16,11 @@ class Operation
 
     # Ignores any attributes and tells us if the tag is the same e.g. <p> and
     # <p style="margin: 2px;"> are the same.
+    # The exception to this rule is our-embed tags, where we
+    # always want a full replacement.
     def same_tag?
+      return false if contains_our_embed_tags?
+
       pattern = /<([^>\s]+)[\s>].*/
       first_tagname = pattern.match(old_text) # nil means they are not tags
       first_tagname = first_tagname[1] if first_tagname
@@ -27,6 +31,10 @@ def same_tag?
       first_tagname && (first_tagname == second_tagname)
     end
 
+    def contains_our_embed_tags?
+      old_text.match?(%r{^<\/?our-embed}) && new_text.match?(%r{^<\/?our-embed})
+    end
+
     def old_text
       old_words.join
     end

lib/htmldiff/word.rb

@@ -26,6 +26,14 @@ def iframe_tag?
       (@word[0..7].downcase =~ %r{^<\/?iframe ?})
     end
 
+    def embed_or_blank_opening_tag?
+      @word =~ Regexp.union(/^<span[^>]*class="[^"]*blank[^"]*"[^>]*>$/i, /^<our-embed[^>]*>$/i)
+    end
+
+    def closed_embed_or_blank_tag?
+      @word =~ /^<(span[^>]*class="[^"]*blank[^"]*"[^>]*|our-embed[^>]*)><\/(span|our-embed)>$/i
+    end
+
     def tag?
       opening_tag? || closing_tag? || standalone_tag?
     end

spec/operation_spec.rb

@@ -24,6 +24,15 @@
       end
     end
 
+    context 'with `our-embed` tags that may have different attributes' do
+      let(:old_tag) { '<our-embed src="gid://123">' }
+      let(:new_tag) { '<our-embed src="gid://456">' }
+
+      it 'returns false for matching and non-matching `our-embed` tags' do
+        expect(operation.same_tag?).to be_false
+      end
+    end
+
     context 'with two different tags' do
       let(:old_tag) { '<p>' }
       let(:new_tag) { '<b>' }