Better handling of NFD chars

fzf

@@ -137,21 +137,21 @@ class FZF
       NFC_END   = NFC_BEGIN + CHOSUNGS * JUNGSUNGS * JONGSUNGS
 
       def self.nfd str
-        ret = ''
-        str.split(//).each do |c|
+        str.split(//).map do |c|
           cp = c.ord
           if cp >= NFC_BEGIN && cp < NFC_END
+            chr = ''
             idx  = cp - NFC_BEGIN
             cho  = CHOSUNG  + idx / JJCOUNT
             jung = JUNGSUNG + (idx % JJCOUNT) / JONGSUNGS
             jong = JONGSUNG + idx % JONGSUNGS
-            ret << cho << jung
-            ret << jong if jong != JONGSUNG
+            chr << cho << jung
+            chr << jong if jong != JONGSUNG
+            chr
           else
-            ret << c
+            c
           end
         end
-        ret
       end
 
       def self.to_nfc arr
@@ -215,16 +215,14 @@ class FZF
 
     class Matcher
       def query_chars q
-        UConv.nfd(q).split(//)
+        UConv.nfd(q)
       end
-    end
-  else
-    module UConv
-      def self.nfd str
-        str
+
+      def sanitize q
+        UConv.nfd(q).join
       end
     end
-
+  else
     def convert_item item
       item
     end
@@ -233,6 +231,10 @@ class FZF
       def query_chars q
         q.split(//)
       end
+
+      def sanitize q
+        q
+      end
     end
   end
 
@@ -755,7 +757,8 @@ class FZF
         q = q.downcase if @rxflag != 0
         Regexp.new(query_chars(q).inject('') { |sum, e|
           e = Regexp.escape e
-          sum << "#{e}[^#{e}]*?"
+          sum << (e.length > 1 ? "(?:#{e}).*?" :  # FIXME: not equivalent
+                                 "#{e}[^#{e}]*?")
         }, @rxflag)
       end
     end
@@ -810,13 +813,13 @@ class FZF
               nil
             when /^'/
               w.length > 1 ?
-                Regexp.new(UConv.nfd(Regexp.escape(w[1..-1])), rxflag) : nil
+                Regexp.new(sanitize(Regexp.escape(w[1..-1])), rxflag) : nil
             when /^\^/
               w.length > 1 ?
-                Regexp.new('^' << UConv.nfd(Regexp.escape(w[1..-1])), rxflag) : nil
+                Regexp.new('^' << sanitize(Regexp.escape(w[1..-1])), rxflag) : nil
             when /\$$/
               w.length > 1 ?
-                Regexp.new(UConv.nfd(Regexp.escape(w[0..-2])) << '$', rxflag) : nil
+                Regexp.new(sanitize(Regexp.escape(w[0..-2])) << '$', rxflag) : nil
             else
               fuzzy_regex w
             end, invert ]

fzf.gemspec

@@ -1,7 +1,7 @@
 # coding: utf-8
 Gem::Specification.new do |spec|
   spec.name          = 'fzf'
-  spec.version       = '0.4.1'
+  spec.version       = '0.4.2'
   spec.authors       = ['Junegunn Choi']
   spec.email         = ['junegunn.c@gmail.com']
   spec.description   = %q{Fuzzy finder for your shell}

test/test_fzf.rb

@@ -293,21 +293,23 @@ class TestFZF < MiniTest::Unit::TestCase
     def test_nfd
       nfc = '한글'
       nfd = FZF::UConv.nfd(nfc)
-      assert_equal 6, nfd.length
-      assert_equal NFD, nfd
+      assert_equal 2, nfd.length
+      assert_equal 6, nfd.join.length
+      assert_equal NFD, nfd.join
     end
 
     def test_nfd_fuzzy_matcher
       matcher = FZF::FuzzyMatcher.new 0
-      match = matcher.match([NFD], '할', '', '')
-      assert_equal [[NFD, [[0, 6]]]], match
-      assert_equal ['한글', [[0, 2]]], FZF::UConv.nfc(*match.first)
+      assert_equal [], matcher.match([NFD + NFD], '할', '', '')
+      match   = matcher.match([NFD + NFD], '글글', '', '')
+      assert_equal [[NFD + NFD, [[3, 12]]]], match
+      assert_equal ['한글한글', [[1, 4]]], FZF::UConv.nfc(*match.first)
     end
 
     def test_nfd_extended_fuzzy_matcher
       matcher = FZF::ExtendedFuzzyMatcher.new 0
-      assert_equal [], matcher.match([NFD], "'할", '', '')
-      match = matcher.match([NFD], "'한글", '', '')
+      assert_equal [], matcher.match([NFD], "'글글", '', '')
+      match   = matcher.match([NFD], "'한글", '', '')
       assert_equal [[NFD, [[0, 6]]]], match
       assert_equal ['한글', [[0, 2]]], FZF::UConv.nfc(*match.first)
     end