With the steps following you should get Ruby/Rails up running with Unicode (UTF8) support. I used MySQL under debian/etch and ferret for full text searching.

Configuring Linux

1. Run ‘dpkg-reconfigure locales’ and be sure to configure de_DE.UTF-8@euro
2. Set de_DE.UTF-8@euro as system-default

Configuring MySQL

1. In /etc/mysql/my.cnf set ‘default-character-set = utf8′ within the [mysqld] section
2. Set ‘default-character-set = utf8′ in [mysql] section also
3. Set Database and Table Collation to ‘utf8_unicode_ci’. (Good introduction is available at julik.textdriven.com and at wiki.rubyonrails.org)

If you need to convert your database tables this article at marstononline.com might be helpful.

Configuring Apache2

Be sure Apache2 is running with ‘AddDefaultCharset utf-8′ in your (virtual-)server. But this should be default anyway.

Configuring your Rails Application

To be sure, set ‘encoding: utf8′ in your config/database.yml for all three databases (development, test and production).

Configuring ferret/acts_as_ferret

I am using ferret with the acts_as_ferret plugin for searching. Be sure to set the following in your config/environment.rb:

# Set this for UTF use (Umlaute)
# http://www.ruby-forum.com/topic/103360#228612
$KCODE = ‘u’
require ‘jcode’

# Ferret search plugin
ENV['LANG'] = ‘de_DE.UTF-8@euro’
ENV['LC_TIME'] = ‘C’
require ‘acts_as_ferret’

This entry was posted on Sunday, April 1st, 2007 at 1:41 pm and is filed under dies und jenes. You can follow any responses to this entry through the RSS 2.0 feed. Responses are currently closed, but you can trackback from your own site.