File: README

Gibberish Rails 0.1

=============

An extension to Gibberish and Rails to use it to create a multilingual web site using the around filter and Gibberish.

The latest version of gibberish_rails can be found here: svn.myutil.com/projects/plugins/gibberish_rails/

Other info can be found here: myutil.com/2008/1/22/gibberish_rails-a-ruby-on-rails-plugin-to-translate-rails-with-gibberish

This is an alpha release. Tests are not written yet. I expected this plugin to be radically rewritten (simplified?) once Rails core is fixed to allow for localization of messages. The issue for Rails core tracking this is here: dev.rubyonrails.org/ticket/9726. I probably didn‘t find and fix all the messages in validates let alone rails but this should be useful to anyone trying to use Gibberish to run a multi-lingual web site.

This plugin:

Adds the method Gibberish.supported? to Gibberish::Localize to use it‘s map of languages to determine if a locale is supported. For example Gibberish.supported?(:en) returns true. It can be passed a string or a symbol. The previous Gibberish API returned an array of languages, but it was just so much more efficient to use the hash Gibberish had already built I added my method to the api. My Helper (6) below needs this method.
Redefines the Rails core method ActionView::Helpers::ActiveRecordHelper.error_message_for so messages are translates with Gibberish.
Overrides default error messages in ActiveReccord:Errors to include a Gibberish key. The key is blank (i.e. the entire string) so they can be automatically re-Gibberished in step 4. (next).
The method "add" in ActiveRecord::Errors has been overridden retranslate messages during the around filter. This become necessary because Rails 2.0.2 caches parts of messages at the class load time.
Redefines the method Validations.ClassMethods.validates_length_of to retranslate the messages it caches at class load up time.
Adds a helper method "GibberishRails.best_guess_locale" that can be used in an around filter to determine the "best guess" locale the application should be using. The highest priority is given to a parameter, followed by HTTP_ACCEPT_LANGUAGE, and finally the Gibberish default. Also, if the user wants es-mx and we support es, it will return a local of es. Here is an example usage that uses the session to cache the locale so if you are using the HTTP_ACCEPT_LANGUAGE (instead of RESTful routes), it will cache in the session instead of parsing headers every time:
```
   class ApplicationController < ActionController::Base
     # ...
     around_filter :use_best_guess_locale
     # ...
     # Set's the best guess language Giberish's around filter
     def use_best_guess_locale
       session[:locale] = GibberishRails.best_guess_locale(params[:locale], request.env['HTTP_ACCEPT_LANGUAGE']) if ( ! session[:locale] || params[:locale] || RAILS_ENV == 'development' )

       Gibberish.use_language(session[:locale]) {
         yield
       }
     end
   end
```

Notes:

This plugin doesn‘t know anything about your models. In prototyping, I translated restful_authentication login and signup screens and need gibberish to pick up some of my model attribute names and some other miscellaneous strings. So, I hackishly put a file in my models directory that‘s only use is for Peter‘s gibberish_translate plugin to pick up the strings:

      -------- models/translate_me.rb------------------
                        "Password confirmation"[]
                        "Password"[]
                        "user"[]

If you are using Peter Marklund‘s gibberish_translate (www.marklunds.com/articles/one/376), the you should change the directories it search to include this one. Also Peter‘s plugin doesn‘t pick up the Gibberish String "foo"[] (which is another way of saying "foo"[:foo]). Redefine Gibberish::Extractor methods in his plugin to be:

      def dirs_to_search
        %w(app config lib vendor/plugins/gibberish_rails).map { |dir| "#{RAILS_ROOT}/#{dir}" }
      end

      def message_pattern(start_token, end_token)
        /#{start_token}((?:[^#{end_token}](?:\\#{end_token})?)+)#{end_token}\[:*([a-z_]*)[,\]]/m
      end

      def add_messages(contents, start_token, end_token)
        contents.scan(message_pattern(start_token, end_token)).each do |text, key|
          key = text.tr('[   ]', '_').downcase if ( key == '' )
          add_message(key, remove_quotes(text, end_token))
        end
      end

Known Issue:

  When the fields in a form are translated and validated, the div's
  that created the red boxes aren't working...have to look into
  that at some point.