Fix adjacent emphasis tags producing invalid markdown

lib/reverse_markdown/cleaner.rb

@@ -5,10 +5,32 @@ def tidy(string)
       result = remove_inner_whitespaces(string)
       result = remove_newlines(result)
       result = remove_leading_newlines(result)
+      result = merge_adjacent_emphasis(result)
       result = clean_tag_borders(result)
       clean_punctuation_characters(result)
     end
 
+    def merge_adjacent_emphasis(string)
+      result = string
+
+      # Merge adjacent underscore emphasis: _X__Y_ → _XY_
+      # Apply repeatedly for multiple adjacent tags
+      loop do
+        new_result = result.gsub(/_([^_\n]+)__([^_\n]+)_/, '_\1\2_')
+        break if new_result == result
+        result = new_result
+      end
+
+      # Merge adjacent strong emphasis: **X****Y** → **XY**
+      loop do
+        new_result = result.gsub(/\*\*([^*\n]+)\*\*\*\*([^*\n]+)\*\*/, '**\1\2**')
+        break if new_result == result
+        result = new_result
+      end
+
+      result
+    end
+
     def remove_newlines(string)
       string.gsub(/\n{3,}/, "\n\n")
     end

spec/lib/reverse_markdown/cleaner_spec.rb

@@ -60,6 +60,43 @@
     end
   end
 
+  describe '#merge_adjacent_emphasis' do
+    it 'merges two adjacent underscore emphasis tags' do
+      result = cleaner.merge_adjacent_emphasis('_a__b_')
+      expect(result).to eq '_ab_'
+    end
+
+    it 'merges three adjacent underscore emphasis tags' do
+      result = cleaner.merge_adjacent_emphasis('_a__b__c_')
+      expect(result).to eq '_abc_'
+    end
+
+    it 'merges two adjacent strong emphasis tags' do
+      result = cleaner.merge_adjacent_emphasis('**a****b**')
+      expect(result).to eq '**ab**'
+    end
+
+    it 'merges three adjacent strong emphasis tags' do
+      result = cleaner.merge_adjacent_emphasis('**a****b****c**')
+      expect(result).to eq '**abc**'
+    end
+
+    it 'does not merge emphasis tags separated by whitespace' do
+      result = cleaner.merge_adjacent_emphasis('_a_ _b_')
+      expect(result).to eq '_a_ _b_'
+    end
+
+    it 'does not merge strong tags separated by whitespace' do
+      result = cleaner.merge_adjacent_emphasis('**a** **b**')
+      expect(result).to eq '**a** **b**'
+    end
+
+    it 'handles mixed content correctly' do
+      result = cleaner.merge_adjacent_emphasis('text _a__b_ more **c****d** end')
+      expect(result).to eq 'text _ab_ more **cd** end'
+    end
+  end
+
   describe '#clean_tag_borders' do
     context 'with default_border is set to space' do
       before { ReverseMarkdown.config.tag_border = ' ' }

spec/lib/reverse_markdown/converters/em_spec.rb

@@ -25,4 +25,18 @@
     expect(result).to include('_hello_')
     expect(result).to include('_world_')
   end
+
+  it 'merges adjacent em tags into single emphasis' do
+    # Issue #99: Adjacent emphasis tags like <em>wo</em><em>rd</em>
+    # should produce _word_ not _wo__rd_
+    expect(ReverseMarkdown.convert('<em>wo</em><em>rd</em>')).to eq '_word_'
+  end
+
+  it 'merges multiple adjacent em tags' do
+    expect(ReverseMarkdown.convert('<em>a</em><em>b</em><em>c</em>')).to eq '_abc_'
+  end
+
+  it 'keeps separate emphasis when tags have whitespace between them' do
+    expect(ReverseMarkdown.convert('<em>a</em> <em>b</em>')).to eq '_a_ _b_'
+  end
 end