From 9ba15e378a601cc15e2625748fb1e8c3750c82b6 Mon Sep 17 00:00:00 2001
From: Lukas Bischof <lukas.bischof@renuo.ch>
Date: Fri, 14 Oct 2022 21:45:09 +0200
Subject: [PATCH] Provide option to strip attributes for empty tags

Introduce strip_attributes_of_empty_tags option
which can be set to have consistent return values for empty tags
by neglecting their attributes.
---
 lib/nori.rb                  | 19 ++++++++-------
 lib/nori/xml_utility_node.rb |  4 ++-
 spec/nori/nori_spec.rb       | 47 ++++++++++++++++++++++++++++++++++++
 3 files changed, 60 insertions(+), 10 deletions(-)

diff --git a/lib/nori.rb b/lib/nori.rb
index ba55215..d404dd2 100644
--- a/lib/nori.rb
+++ b/lib/nori.rb
@@ -15,15 +15,16 @@ def self.hash_key(name, options = {})
 
   def initialize(options = {})
     defaults = {
-      :strip_namespaces              => false,
-      :delete_namespace_attributes   => false,
-      :convert_tags_to               => nil,
-      :convert_attributes_to         => nil,
-      :empty_tag_value               => nil,
-      :advanced_typecasting          => true,
-      :convert_dashes_to_underscores => true,
-      :scrub_xml                     => true,
-      :parser                        => :nokogiri
+      :strip_namespaces               => false,
+      :delete_namespace_attributes    => false,
+      :convert_tags_to                => nil,
+      :convert_attributes_to          => nil,
+      :empty_tag_value                => nil,
+      :advanced_typecasting           => true,
+      :convert_dashes_to_underscores  => true,
+      :scrub_xml                      => true,
+      :strip_attributes_of_empty_tags => false,
+      :parser                         => :nokogiri
     }
 
     validate_options! defaults.keys, options.keys
diff --git a/lib/nori/xml_utility_node.rb b/lib/nori/xml_utility_node.rb
index 63d74a4..cf55ff4 100644
--- a/lib/nori/xml_utility_node.rb
+++ b/lib/nori/xml_utility_node.rb
@@ -170,7 +170,9 @@ def to_hash
               out.merge!( k => v.map{|e| e.to_hash[k]})
             end
           end
-          out.merge! prefixed_attributes unless attributes.empty?
+          if !(@options[:strip_attributes_of_empty_tags] && @children.empty?) && attributes.any?
+            out.merge! prefixed_attributes
+          end
           out = out.empty? ? @options[:empty_tag_value] : out
         end
 
diff --git a/spec/nori/nori_spec.rb b/spec/nori/nori_spec.rb
index e673722..d87524f 100644
--- a/spec/nori/nori_spec.rb
+++ b/spec/nori/nori_spec.rb
@@ -640,6 +640,53 @@
         expect(parse(' ')).to eq({})
       end
 
+      context "with strip_attributes_of_empty_tags set to true" do
+        it "should strip attributes of empty tags" do
+          expect(parse('<foo bar="" />', strip_attributes_of_empty_tags: true)).to eq({ 'foo' => nil })
+        end
+
+        it "should strip attributes only of empty tags also in a nested structure" do
+          xml = <<-XML
+            <content>
+              <foo />
+              <foo bar="baz" />
+              <foo bar="baz" bar2="baz2"></foo>
+              <foo bar="baz">actual content</foo>
+            </content>
+          XML
+
+          expect(parse(xml, strip_attributes_of_empty_tags: true)).to match({
+            "content" => {
+              "foo" => [nil, nil, nil, be_a(Nori::StringWithAttributes).and(eq("actual content"))],
+            }
+          })
+        end
+      end
+
+      context "with strip_attributes_of_empty_tags set to false" do
+        it "should not strip any attributes" do
+          xml = <<-XML
+            <content>
+              <foo />
+              <foo bar="baz" />
+              <foo bar="baz" bar2="baz2"></foo>
+              <foo bar="baz">actual content</foo>
+            </content>
+          XML
+
+          expect(parse(xml, strip_attributes_of_empty_tags: false)).to match({
+            "content" => {
+              "foo" => [
+                nil,
+                {"@bar"=>"baz"},
+                {"@bar"=>"baz", "@bar2"=>"baz2"},
+                be_a(Nori::StringWithAttributes).and(eq("actual content"))
+              ],
+            }
+          })
+        end
+      end
+
     end
   end