Pull in latest multibyte patch. Closes #6346 [Manfred Stienstra]

git-svn-id: http://svn-commit.rubyonrails.org/rails/trunk@5224 5ecf4fe2-1ee6-0310-87b1-e25e094e27de

activesupport/CHANGELOG

@@ -1,6 +1,8 @@
 *SVN*
 
-* Add ActiveSupport::Multibyte.  Provides String#chars which lets you deal with strings as a sequence of chars, not of bytes. Closes #6242 [Julian Tarkhanov, Manfred Stienstra & Jan Behrens]
+* Pull in latest multibye patch. Closes #6346 [Manfred Stienstra]
+
+* Add ActiveSupport::Multibyte.  Provides String#chars which lets you deal with strings as a sequence of chars, not of bytes. Closes #6242 [Julian Tarkhanov, Manfred Stienstra, Thijs van der Vossen & Jan Behrens]
 
 * Fix issue with #class_inheritable_accessor saving updates to the parent class when initialized with an Array or Hash [mojombo]
 

activesupport/lib/active_support/multibyte/generators/generate_tables.rb

@@ -18,7 +18,8 @@ module ActiveSupport::Multibyte::Handlers #:nodoc:
     SOURCES = {
       :codepoints => BASE_URI + 'UnicodeData.txt',
       :composition_exclusion => BASE_URI + 'CompositionExclusions.txt',
-      :grapheme_break_property => BASE_URI + 'auxiliary/GraphemeBreakProperty.txt'
+      :grapheme_break_property => BASE_URI + 'auxiliary/GraphemeBreakProperty.txt',
+      :cp1252 => 'http://unicode.org/Public/MAPPINGS/VENDORS/MICSFT/WINDOWS/CP1252.TXT'
     }
     
     def initialize
@@ -33,6 +34,7 @@ module ActiveSupport::Multibyte::Handlers #:nodoc:
       @ucd.composition_exclusion = []
       @ucd.composition_map = {}
       @ucd.boundary = {}
+      @ucd.cp1252 = {}
     end
     
     def parse_codepoints(line)
@@ -87,6 +89,12 @@ module ActiveSupport::Multibyte::Handlers #:nodoc:
       end
     end
     
+    def parse_cp1252(line)
+      if line =~ /^([0-9A-Fx]+)\s([0-9A-Fx]+)/i
+        @ucd.cp1252[$1.hex] = $2.hex
+      end
+    end
+    
     def create_composition_map
       @ucd.codepoints.each do |_, cp|
         if !cp.nil? and cp.combining_class == 0 and cp.decomp_type.nil? and !cp.decomp_mapping.nil? and cp.decomp_mapping.length == 2 and @ucd[cp.decomp_mapping[0]].combining_class == 0 and !@ucd.composition_exclusion.include?(cp.code)
@@ -125,7 +133,7 @@ module ActiveSupport::Multibyte::Handlers #:nodoc:
     
     def dump_to(filename)
       File.open(filename, 'wb') do |f|
-        f.write Marshal.dump([@ucd.codepoints, @ucd.composition_exclusion, @ucd.composition_map, @ucd.boundary])
+        f.write Marshal.dump([@ucd.codepoints, @ucd.composition_exclusion, @ucd.composition_map, @ucd.boundary, @ucd.cp1252])
       end
     end
   end

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

@@ -7,12 +7,12 @@ module ActiveSupport::Multibyte::Handlers
   end
   
   class UnicodeDatabase #:nodoc:
-    attr_accessor :codepoints, :composition_exclusion, :composition_map, :boundary
+    attr_accessor :codepoints, :composition_exclusion, :composition_map, :boundary, :cp1252
     
     # Creates a new UnicodeDatabase instance and loads the database.
     def initialize
       begin
-        @codepoints, @composition_exclusion, @composition_map, @boundary = self.class.load
+        @codepoints, @composition_exclusion, @composition_map, @boundary, @cp1252 = self.class.load
       rescue Exception => e
           raise IOError.new("Couldn't load the unicode tables for UTF8Handler (#{e.message}), handler is unusable")
       end
@@ -20,6 +20,7 @@ module ActiveSupport::Multibyte::Handlers
       @composition_exclusion ||= []
       @composition_map ||= {}
       @boundary ||= {}
+      @cp1252 ||= {}
       
       # Redefine the === method so we can write shorter rules for grapheme cluster breaks
       @boundary.each do |k,_|
@@ -41,26 +42,11 @@ module ActiveSupport::Multibyte::Handlers
     
     # Returns the filename for the data file for this version
     def self.filename
-      File.expand_path File.join(dirname, "unicode_tables-#{VERSION}.dat")
+      File.expand_path File.join(dirname, "unicode_tables.dat")
     end
     
     # Loads the unicode database and returns all the internal objects of UnicodeDatabase
     def self.load
-      begin
-        return load_file(filename)
-      rescue Exception
-        # If we can't load our own version, try the rest
-        Dir["#{dirname}/*.dat"].sort.each do |dat|
-          begin
-            return load_file(dat)
-          rescue Exception
-          end
-        end
-      end
-      raise IOError.new("Can't load a marshal file for your version of Ruby")
-    end
-    
-    def self.load_file(filename)
       File.open(self.filename, 'rb') { |f| Marshal.load f.read }
     end
   end
@@ -275,7 +261,11 @@ module ActiveSupport::Multibyte::Handlers
       
       # Strips all the non-utf-8 bytes from the string resulting in a valid utf-8 string
       def tidy_bytes(str)
-        str.split(//u).reject { |c| !UTF8_PAT.match(c) }.join
+        str.unpack('C*').map { |n|
+          n < 128 ? n.chr :
+          n < 160 ? [UCD.cp1252[n] || n].pack('U') :
+          n < 192 ? "\xC2" + n.chr : "\xC3" + (n-64).chr
+        }.join
       end
       
       protected

activesupport/test/multibyte_chars_test.rb

@@ -139,14 +139,15 @@ class CharsTest < Test::Unit::TestCase
   
   def test_resilience
     assert_nothing_raised do
-      assert_equal 1, @s[:bytes].chars.size, "There's only one valid utf-8 byte in the string"
+      assert_equal 5, @s[:bytes].chars.size, "The sequence contains five interpretable bytes"
     end
+    reversed = [0xb8, 0x17e, 0x8, 0x2c6, 0xa5].reverse.pack('U*')
     assert_nothing_raised do
-      assert_equal "\010", @s[:bytes].chars.reverse, "There's only one valid utf-8 byte in the string"
+      assert_equal reversed, @s[:bytes].chars.reverse.to_s, "Reversing the string should only yield interpretable bytes"
     end
     assert_nothing_raised do
       @s[:bytes].chars.reverse!
-      assert_equal "\010", @s[:bytes], "There's only one valid utf-8 byte in the string"
+      assert_equal reversed, @s[:bytes].to_s, "Reversing the string should only yield interpretable bytes"
     end
   end
   

activesupport/test/multibyte_handler_test.rb

@@ -224,9 +224,17 @@ module UTF8HandlingTest
   end
   
   def test_tidy_bytes
-    assert_equal "\010", @handler.tidy_bytes(@bytestring)
-    assert_equal "a\010a", @handler.tidy_bytes('a' + @bytestring + 'a')
+    result = [0xb8, 0x17e, 0x8, 0x2c6, 0xa5].pack('U*')
+    assert_equal result, @handler.tidy_bytes(@bytestring)
+    assert_equal "a#{result}a", @handler.tidy_bytes('a' + @bytestring + 'a')
     assert_nothing_raised { @handler.tidy_bytes(@bytestring).unpack('U*') }
+    
+    assert_equal "\xC3\xA7", @handler.tidy_bytes("\xE7") # iso_8859_1: small c cedilla
+    assert_equal "\xC2\xA9", @handler.tidy_bytes("\xA9") # iso_8859_1: copyright symbol
+    assert_equal "\xE2\x80\x9C", @handler.tidy_bytes("\x93") # win_1252: left smart quote
+    assert_equal "\xE2\x82\xAC", @handler.tidy_bytes("\x80") # win_1252: euro
+    assert_equal "\x00", @handler.tidy_bytes("\x00") # null char
+    assert_equal [0xef, 0xbf, 0xbd].pack('U*'), @handler.tidy_bytes("\xef\xbf\xbd") # invalid char
   end
   
   protected