Question

I’m having a problem that’s testing my sanity. In my rails application, I’m loading a escaped UTF-8 string from an I18n translation file to be used in my application routes. However, rails is not unescaping the string when used in routing. Here’s my translation file:

--- 
ru: 
  activerecord: 
    models: 
      item: 
        other: "\xD0\xA2\xD0\xBE\xD0\xB2\xD0\xB0\xD1\x80\xD1\x8B" # UTF-8 escaped version of "Товары". (to_yaml is doing the escaping btw)

And here’s my config/routes.rb

match "/#{Item.model_name.human(:count => :other).downcase}" => "items#index"

However, when navigating to this route, I get:

Routing Error

No route matches "/%d0%a2%d0%be%d0%b2%d0%b0%d1%80%d1%8b"

If the UTF-8 is stored in the translation file unescaped, everything works fine. to_yaml is escaping the string and causing the problem. Is there any way to force yaml to retain the original string?

BTW, I’m using Rails 3.0.7, Ruby 1.9.2. Thanks guys!

Answer 1

I was able to store the original UTF-8 string without converting the string to Ruby internal encoding by using ya2yaml instead of to_yaml. However, I had to encode the params hash’s keys & values to UTF-8(keys were being encoded as ASCII-8BIT and values as UTF-8) before the yaml was generated properly:

  def utf8_hash(some_hash) # convert hash key & values to utf-8 for proper translation
    new_hash = Hash.new
    some_hash.each do |key, value|
      new_hash[key.encode(Encoding::UTF_8)] = value.to_s.encode(Encoding::UTF_8)
    end    
    new_hash
  end

  utf8_hash(params).ya2yaml