Ticket #9165: add_rjust_ljust_and_center_to_utf8_handler.diff

activesupport/test/multibyte_handler_test.rb

@@ -233 +233 @@
-    assert_equal "Κλη αααα!", s
-  end
-  
+  def test_rjust
+    s = "Καη"
+    assert_raises(ArgumentError) { @handler.rjust(s, 10, '') }
+    assert_raises(ArgumentError) { @handler.rjust(s) }
+    assert_equal "Καη", @handler.rjust(s, -3)
+    assert_equal "Καη", @handler.rjust(s, 0)
+    assert_equal "Καη", @handler.rjust(s, 3)
+    assert_equal "  Καη", @handler.rjust(s, 5)
+    assert_equal "    Καη", @handler.rjust(s, 7)
+    assert_equal "----Καη", @handler.rjust(s, 7, '-')
+    assert_equal "ααααΚαη", @handler.rjust(s, 7, 'α')
+    assert_equal "abaΚαη", @handler.rjust(s, 6, 'ab')
+    assert_equal "αηαΚαη", @handler.rjust(s, 6, 'αη')
+  end
+  
+  def test_ljust
+    s = "Καη"
+    assert_raises(ArgumentError) { @handler.ljust(s, 10, '') }
+    assert_raises(ArgumentError) { @handler.ljust(s) }
+    assert_equal "Καη", @handler.ljust(s, -3)
+    assert_equal "Καη", @handler.ljust(s, 0)
+    assert_equal "Καη", @handler.ljust(s, 3)
+    assert_equal "Καη  ", @handler.ljust(s, 5)
+    assert_equal "Καη    ", @handler.ljust(s, 7)
+    assert_equal "Καη----", @handler.ljust(s, 7, '-')
+    assert_equal "Καηαααα", @handler.ljust(s, 7, 'α')
+    assert_equal "Καηaba", @handler.ljust(s, 6, 'ab')
+    assert_equal "Καηαηα", @handler.ljust(s, 6, 'αη')
+  end
+  
+  def test_center
+    s = "Καη"
+    assert_raises(ArgumentError) { @handler.center(s, 10, '') }
+    assert_raises(ArgumentError) { @handler.center(s) }
+    assert_equal "Καη", @handler.center(s, -3)
+    assert_equal "Καη", @handler.center(s, 0)
+    assert_equal "Καη", @handler.center(s, 3)
+    assert_equal "Καη ", @handler.center(s, 4)
+    assert_equal " Καη ", @handler.center(s, 5)
+    assert_equal " Καη  ", @handler.center(s, 6)
+    assert_equal "--Καη--", @handler.center(s, 7, '-')
+    assert_equal "--Καη---", @handler.center(s, 8, '-')
+    assert_equal "ααΚαηαα", @handler.center(s, 7, 'α')
+    assert_equal "ααΚαηααα", @handler.center(s, 8, 'α')
+    assert_equal "aΚαηab", @handler.center(s, 6, 'ab')
+    assert_equal "abΚαηab", @handler.center(s, 7, 'ab')
+    assert_equal "ababΚαηabab", @handler.center(s, 11, 'ab')
+    assert_equal "αΚαηαη", @handler.center(s, 6, 'αη')
+    assert_equal "αηΚαηαη", @handler.center(s, 7, 'αη')
+  end
+  
-  def test_strip
-    # A unicode aware version of strip should strip all 26 types of whitespace. This includes the NO BREAK SPACE
-    # aka BOM (byte order mark). The byte order mark has no place in UTF-8 because it's used to detect LE and BE.

activesupport/lib/active_support/multibyte/handlers/utf8_handler.rb

@@ -178 +178 @@
-        str.replace(result.pack('U*'))
-      end
-      
+      # Works just like String#rjust, only integer specifies characters instead of bytes.
+      #
+      # Example:
+      #
+      #   "Ÿ cup".chars.rjust(8).to_s
+      #   #=> "   Ÿ cup"
+      #
+      #   "Ÿ cup".chars.rjust(8, " ").to_s # Use non-breaking whitespace
+      #   #=> "   Ÿ cup"
+      def rjust(str, integer, padstr=' ')
+        justify(str, integer, :right, padstr)
+      end
+      
+      # Works just like String#ljust, only integer specifies characters instead of bytes.
+      #
+      # Example:
+      #
+      #   "Ÿ cup".chars.rjust(8).to_s
+      #   #=> "Ÿ cup   "
+      #
+      #   "Ÿ cup".chars.rjust(8, " ").to_s # Use non-breaking whitespace
+      #   #=> "Ÿ cup   "
+      def ljust(str, integer, padstr=' ')
+        justify(str, integer, :left, padstr)
+      end
+      
+      # Works just like String#center, only integer specifies characters instead of bytes.
+      #
+      # Example:
+      #
+      #   "Ÿ cup".chars.center(8).to_s
+      #   #=> " Ÿ cup  "
+      #
+      #   "Ÿ cup".chars.center(8, " ").to_s # Use non-breaking whitespace
+      #   #=> " Ÿ cup  "
+      def center(str, integer, padstr=' ')
+        justify(str, integer, :center, padstr)
+      end
+      
-      # Does Unicode-aware rstrip
-      def rstrip(str)
-        str.gsub(UNICODE_TRAILERS_PAT, '')
@@ -380 +419 @@
-        unpacked.flatten
-      end
-      
+      # Justifies a string in a certain way. Valid values for <tt>way</tt> are <tt>:right</tt>, <tt>:left</tt> and
+      # <tt>:center</tt>. Is primarily used as a helper method by <tt>rjust</tt>, <tt>ljust</tt> and <tt>center</tt>.
+      def justify(str, integer, way, padstr=' ')
+        raise ArgumentError, "zero width padding" if padstr.length == 0
+        padsize = integer - size(str)
+        padsize = padsize > 0 ? padsize : 0
+        case way
+        when :right
+          str.dup.insert(0, padding(padsize, padstr))
+        when :left
+          str.dup.insert(-1, padding(padsize, padstr))
+        when :center
+          lpad = padding((padsize / 2.0).floor, padstr)
+          rpad = padding((padsize / 2.0).ceil, padstr)
+          str.dup.insert(0, lpad).insert(-1, rpad)
+        end
+      end
+      
+      # Generates a padding string of a certain size.
+      def padding(padsize, padstr=' ')
+        if padsize != 0
+          slice(padstr * ((padsize / size(padstr)) + 1), 0, padsize)
+        else
+          ''
+        end
+      end
+      
-      # Convert characters to a different case
-      def to_case(way, str)
-        u_unpack(str).map do |codepoint|