char_converter.rb

# encoding: UTF-8
require 'iconv'
require 'uri'

module Support
  class CharConverter

    BAD_GA_ENCODINGS = {
      "%u00DF" => "ß",
      "%u00E4" => "ä",
      "%u00F6" => "ö",
      "%u00FC" => "ü",
      "%uFFFD" => "ä"
    }

    CONVERTER = Iconv.new('UTF-8', 'windows-1252')

    def initialize(app)
      @app = app
    end

    def call(env)
      @app.call(self.class.kill_windows_encoding(env))
    end

    def self.kill_windows_encoding(env)
      ["QUERY_STRING", "REQUEST_URI", "HTTP_REFERER", "HTTP_COOKIE"].each do |key|
        next unless value = env[key]
        value = kill_windows_encoding_for_string(value)
        value = kill_ga_windows_encoding_for_string(value)
        env[key] = value
      end
      env
    end

    def self.kill_ga_windows_encoding_for_string(string)
      # bad cookie encodings kill rack https://github.com/rack/rack/issues/225
      BAD_GA_ENCODINGS.inject(string) { |s, entry| s.gsub(/#{entry.first}/, entry.last) }
    end

    def self.kill_windows_encoding_for_string(string)
      return nil unless string
      # decode URI from %E4 to \xE4
      decoded = URI.decode(string)
      # force UTF-8 cause it usally works within ascii 8-bit
      unless decoded.force_encoding('UTF-8').valid_encoding?
        begin
          # wild guess that it's a fucking windows encoding
          string = URI.encode(CONVERTER.iconv(decoded))
        rescue
          # we don't care
          puts $!
        end
      end
      string
    end

  end
end