From b9379e7a26bce4ba76108da1b6f13ea22e236632 Mon Sep 17 00:00:00 2001
From: Go MAEDA <maeda@farend.jp>
Date: Thu, 3 Nov 2022 12:43:28 +0000
Subject: [PATCH] Allow using ideographic space (U+3000) as a separator for
 search terms (#37878).

Patch by Go MAEDA.


git-svn-id: https://svn.redmine.org/redmine/trunk@21952 e93f8b46-1217-0410-a6f0-8f06a7374b81
---
 lib/redmine/search.rb                | 2 +-
 test/unit/lib/redmine/search_test.rb | 6 ++++++
 2 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/lib/redmine/search.rb b/lib/redmine/search.rb
index bf8b7633a..3c1e0700c 100644
--- a/lib/redmine/search.rb
+++ b/lib/redmine/search.rb
@@ -135,7 +135,7 @@ module Redmine
       def tokens
         # extract tokens from the question
         # eg. hello "bye bye" => ["hello", "bye bye"]
-        tokens = @question.scan(%r{((\s|^)"[^"]+"(\s|$)|\S+)}).collect {|m| m.first.gsub(%r{(^\s*"\s*|\s*"\s*$)}, '')}
+        tokens = @question.scan(%r{(([[:space:]]|^)"[^"]+"([[:space:]]|$)|[[:^space:]]+)}).collect {|m| m.first.gsub(%r{(^[[:space:]]*"[[:space:]]*|[[:space:]]*"[[:space:]]*$)}, '')}
         # tokens must be at least 2 characters long
         # but for Chinese characters (Chinese HANZI/Japanese KANJI), tokens can be one character
         # no more than 5 tokens to search for
diff --git a/test/unit/lib/redmine/search_test.rb b/test/unit/lib/redmine/search_test.rb
index e2944b4ac..219f40cfd 100644
--- a/test/unit/lib/redmine/search_test.rb
+++ b/test/unit/lib/redmine/search_test.rb
@@ -24,4 +24,10 @@ class Redmine::Search::Tokenize < ActiveSupport::TestCase
     value = "hello \"bye bye\""
     assert_equal ["hello", "bye bye"], Redmine::Search::Tokenizer.new(value).tokens
   end
+
+  def test_tokenize_should_consider_ideographic_space_as_separator
+    # U+3000 is an ideographic space ("　")
+    value = "全角\u3000スペース"
+    assert_equal %w[全角 スペース], Redmine::Search::Tokenizer.new(value).tokens
+  end
 end