Scrub invalid characters from source XML

Nori::Parser has a new option, :scrub_xml, which defaults to true, when
it's true, the parser will clean invalid or undefined characters from
the string using String#scrub if it's available or String#encode
otherwise.

lib/nori.rb

@@ -22,6 +22,7 @@ def initialize(options = {})
       :empty_tag_value               => nil,
       :advanced_typecasting          => true,
       :convert_dashes_to_underscores => true,
+      :scrub_xml                     => true,
       :parser                        => :nokogiri
     }
 
@@ -40,7 +41,7 @@ def find(hash, *path)
   end
 
   def parse(xml)
-    cleaned_xml = xml.strip
+    cleaned_xml = scrub_xml(xml).strip
     return {} if cleaned_xml.empty?
 
     parser = load_parser @options[:parser]
@@ -77,4 +78,22 @@ def find_value(hash, key)
     nil
   end
 
+  def scrub_xml(string)
+    if @options[:scrub_xml]
+      if string.respond_to? :scrub
+        string.scrub
+      else
+        if string.valid_encoding?
+          string
+        else
+          enc = string.encoding
+          mid_enc = ([:UTF_8, :UTF_16].map { |e| Encoding.find(e) } - [enc]).first
+          string.encode(mid_enc, undef: :replace, invalid: :replace).encode(enc)
+        end
+      end
+    else
+      string 
+    end
+  end
+
 end

spec/nori/api_spec.rb

@@ -108,13 +108,6 @@
       expect(Nori::Parser::Nokogiri).to receive(:parse).and_return({})
       nori.parse("<any>thing</any>")
     end
-
-    it "strips the XML" do
-      xml = double("xml")
-      expect(xml).to receive(:strip).and_return("<any>thing</any>")
-
-      expect(nori.parse(xml)).to eq({ "any" => "thing" })
-    end
   end
 
   context "#parse without :advanced_typecasting" do

spec/nori/nori_spec.rb

@@ -28,6 +28,11 @@
         expect(parse(xml)["tag"].strip).to eq("text inside cdata")
       end
 
+      it "should scrub bad characters" do
+        xml = "<tag>a\xfbc</tag>".force_encoding('UTF-8')
+        expect(parse(xml)["tag"]).to eq("a\uFFFDc")
+      end
+
       it "should transform a simple tag with attributes" do
         xml = "<tag attr1='1' attr2='2'></tag>"
         hash = { 'tag' => { '@attr1' => '1', '@attr2' => '2' } }