Switch slugify regex to support more Unicode character groups (#8167)

Merge pull request 8167
jekyll · May 14, 2020 · f8286b6 · f8286b6
1 parent 11dd893
commit f8286b6
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 2 deletions.
diff --git a/lib/jekyll/utils.rb b/lib/jekyll/utils.rb
@@ -13,8 +13,8 @@ module Utils
     # Constants for use in #slugify
     SLUGIFY_MODES = %w(raw default pretty ascii latin).freeze
     SLUGIFY_RAW_REGEXP = Regexp.new('\\s+').freeze
-    SLUGIFY_DEFAULT_REGEXP = Regexp.new("[^[:alnum:]]+").freeze
-    SLUGIFY_PRETTY_REGEXP = Regexp.new("[^[:alnum:]._~!$&'()+,;=@]+").freeze
+    SLUGIFY_DEFAULT_REGEXP = Regexp.new("[^\\p{M}\\p{L}\\p{Nd}]+").freeze
+    SLUGIFY_PRETTY_REGEXP = Regexp.new("[^\\p{M}\\p{L}\\p{Nd}._~!$&'()+,;=@]+").freeze
     SLUGIFY_ASCII_REGEXP = Regexp.new("[^[A-Za-z0-9]]+").freeze
 
     # Takes a slug and turns it into a simple title.

diff --git a/test/test_utils.rb b/test/test_utils.rb
@@ -176,6 +176,11 @@ class TestUtils < JekyllUnitTest
       assert_equal "5時-6時-三-一四", Utils.slugify("5時〜6時 三・一四")
     end
 
+    should "not replace Unicode 'Mark', 'Letter', or 'Number: Decimal Digit' category characters" do
+      assert_equal "மல்லிப்பூ-வகைகள்", Utils.slugify("மல்லிப்பூ வகைகள்")
+      assert_equal "மல்லிப்பூ-வகைகள்", Utils.slugify("மல்லிப்பூ வகைகள்", :mode => "pretty")
+    end
+
     should "not modify the original string" do
       title = "Quick-start guide"
       Utils.slugify(title)