If UTF-8 encoding isn't valid for InfoExtractor, force the encoding.

Author: knowtheory

lib/docsplit/info_extractor.rb

@@ -27,7 +27,7 @@ def extract_all(pdfs, opts)
       raise ExtractionFailed, result if $? != 0
       # ruby  1.8 (iconv) and 1.9 (String#encode) :
       if String.method_defined?(:encode)
-        result.encode!('UTF-8', 'UTF-8', :invalid => :replace)
+        result.encode!('UTF-8', 'binary', :invalid => :replace, :undef => :replace, :replace => "") unless result.valid_encoding?
       else
         require 'iconv' unless defined?(Iconv)
         ic = Iconv.new('UTF-8//IGNORE','UTF-8')