Simplify EncodedString conversions; fix ArgumentError raised from #split

Also, make Rubocop happy

Author: bf4

lib/rspec/support/encoded_string.rb

@@ -2,8 +2,6 @@ module RSpec
   module Support
     # @private
     class EncodedString
-      MRI_UNICODE_UNKOWN_CHARACTER = "\xEF\xBF\xBD"
-
       def initialize(string, encoding=nil)
         @encoding = encoding
         @source_encoding = detect_source_encoding(string)
@@ -34,19 +32,18 @@ def to_s
         private
 
         def matching_encoding(string)
-          string.encode(@encoding)
-        rescue Encoding::UndefinedConversionError, Encoding::InvalidByteSequenceError
-          normalize_missing(string.encode(@encoding, :invalid => :replace, :undef => :replace))
+          # Converting it to a higher higher character set (UTF-16) and then
+          # back (to UTF-8) ensures that you will strip away invalid or undefined byte sequences.
+          string.
+            encode(::Encoding::UTF_16LE,
+                   :invalid => :replace, :undef => :replace, :replace => '?').
+            encode(@encoding)
+        rescue Encoding::UndefinedConversionError, Encoding::InvalidByteSequenceError, Encoding::CompatibilityError
+          string.encode(@encoding,
+                   :invalid => :replace, :undef => :replace, :replace => '?')
         rescue Encoding::ConverterNotFoundError
-          normalize_missing(string.force_encoding(@encoding).encode(:invalid => :replace))
-        end
-
-        def normalize_missing(string)
-          if @encoding.to_s == "UTF-8"
-            string.gsub(MRI_UNICODE_UNKOWN_CHARACTER.force_encoding(@encoding), "?")
-          else
-            string
-          end
+          string.force_encoding(@encoding).encode(
+                   :invalid => :replace, :replace => '?')
         end
 
         def detect_source_encoding(string)

spec/rspec/support/encoded_string_spec.rb

@@ -105,9 +105,7 @@ module RSpec::Support
         it 'fails to split a string with an invalid byte sequence' do
           message_with_invalid_byte_sequence = "\xEF \255 \xAD I have bad bytes".force_encoding(target_encoding)
           resulting_string = build_encoded_string(message_with_invalid_byte_sequence, target_encoding)
-          expect {
-            resulting_string.split("\n")
-          }.to raise_error(ArgumentError)
+          expect(resulting_string.split("\n")).to eq(["? ? ? I have bad bytes"])
         end
 
         it 'splits the string based on the delimiter accounting for encoding' do