Improved support for non Latin characters in category names. Fixes #128

3 files changed, 8 insertions, 4 deletions

diff --git a/Gemfile b/Gemfile
index ffb614f2..14ee9669 100644
--- a/Gemfile
+++ b/Gemfile
@@ -10,4 +10,5 @@ gem 'haml', '>= 3.1'
 gem 'compass', '>= 0.11'
 gem 'rubypants'
 gem 'rb-fsevent'
-gem 'stringex'
\ No newline at end of file
+gem 'stringex'
+gem 'unicode_utils'
\ No newline at end of file
diff --git a/Gemfile.lock b/Gemfile.lock
index f2fc737d..000eec9b 100644
--- a/Gemfile.lock
+++ b/Gemfile.lock
@@ -42,6 +42,7 @@ GEM
     sass (3.1.7)
     stringex (1.3.0)
     syntax (1.0.0)
+    unicode_utils (1.0.0)
 
 PLATFORMS
   ruby
@@ -58,3 +59,4 @@ DEPENDENCIES
   rdiscount
   rubypants
   stringex
+  unicode_utils
diff --git a/plugins/category_generator.rb b/plugins/category_generator.rb
index aa1180e1..97fb9f30 100644
--- a/plugins/category_generator.rb
+++ b/plugins/category_generator.rb
@@ -16,8 +16,9 @@
 # - category_dir:          The subfolder to build category pages in (default is 'categories').
 # - category_title_prefix: The string used before the category name in the page title (default is
 #                          'Category: ').
-module Jekyll
+require "unicode_utils"
 
+module Jekyll
 
   # The CategoryIndex class creates a single category page for the specified category.
   class CategoryIndex < Page
@@ -68,7 +69,7 @@ module Jekyll
       if self.layouts.key? 'category_index'
         dir = self.config['category_dir'] || 'categories'
         self.categories.keys.each do |category|
-          self.write_category_index(File.join(dir, category.gsub(/_|\W/, '-')), category)
+          self.write_category_index(File.join(dir, UnicodeUtils.nfkd(category).gsub(/[^\x00-\x7F]/, '').gsub(/_|\W/, '-').to_s), category)
         end
 
       # Throw an exception if the layout couldn't be found.
@@ -105,7 +106,7 @@ module Jekyll
     def category_links(categories)
       dir = @context.registers[:site].config['category_dir']
       categories = categories.sort!.map do |item|
-        "<a class='category' href='/#{dir}/#{item.gsub(/_|\W/, '-')}/'>#{item}</a>"
+        "<a class='category' href='/#{dir}/#{UnicodeUtils.nfkd(item).gsub(/[^\x00-\x7F]/, '').gsub(/_|\W/, '-').to_s}/'>#{item}</a>"
       end
 
       case categories.length