Use charlock_holmes to do encoding detection. In my tests it has properly identified incorrect encodings that used to be present on older commits. This will help ensure this won't happen again, giving people instant feedback and allowing all pull requests to be checked

Gemfile

@@ -1,3 +1,4 @@
 group :test do
     gem 'rake'
+    gem 'charlock_holmes'
 end

Rakefile

@@ -2,6 +2,10 @@ task default: %w[test]
 
 task :test do
   Dir["./tests/*.rb"].each do |test_file|
+    begin
     	ruby test_file
+    rescue
+      puts "FAILED #{test_file}!"
+    end
   end
 end

tests/encoding.rb

@@ -1,14 +1,18 @@
 #!/usr/bin/env ruby
+require 'charlock_holmes'
 $file_count = 0;
 markdown_files = Dir["./**/*.html.markdown"]
 markdown_files.each do |file|
   begin
-    file_bin = File.open(file, "rb")
+    contents = File.read(file)
-    contents = file_bin.read
+    detection = CharlockHolmes::EncodingDetector.detect(contents)
-    if ! contents.valid_encoding?
+    case detection[:encoding]
-      puts "#{file} has an invalid encoding! Please save the file in UTF-8!"
+    when 'UTF-8'
-    else
       $file_count = $file_count + 1
+    when 'ISO-8859-1'
+      $file_count = $file_count + 1
+    else
+      puts "#{file} was detected as #{detection[:encoding]} encoding! Please save the file in UTF-8!"
     end
   rescue Exception => msg
     puts msg
@@ -20,6 +24,6 @@ if files_failed != 0
   puts "Please resave the file as UTF-8."
   exit 1
 else
-  puts "Success. All #{$file_count} files passed UTF-8 validity checks"
+  puts "Success. All #{$file_count} files Ruby's UTF-8 validity checks. This won't catch most problems."
   exit 0
 end