Bug #10418: REXML's encoding is broken if reading UTF-16 XML and Encondig.default_internal is set

• Status: Closed
• Assignee: kou (Kouhei Sutou)
• Target version: 2.2.0
• ruby -v: -
• Backport: 2.0.0: UNKNOWN, 2.1: UNKNOWN

[usa (Usaku NAKAMURA)]

Encoding.default_internalがセットされている状態でREXMLにUTF-16なIOを食わせると、REXML::Document#encodingがUTF-16ではなくEncoding.default_internalになります。
以下パッチ。

Index: lib/rexml/source.rb
===================================================================
--- lib/rexml/source.rb	(revision 48095)
+++ lib/rexml/source.rb	(working copy)
@@ -285,7 +285,7 @@
       case @encoding
       when "UTF-16BE", "UTF-16LE"
         @source.binmode
-        @source.set_encoding(@encoding)
+        @source.set_encoding(@encoding, @encoding)
       end
       @line_break = encode(">")
       @pending_buffer, @buffer = @buffer, ""
Index: test/rexml/test_encoding.rb
===================================================================
--- test/rexml/test_encoding.rb	(revision 48095)
+++ test/rexml/test_encoding.rb	(working copy)
@@ -91,8 +91,18 @@
       utf16 = File.open(fixture_path("ticket_110_utf16.xml")) do |f|
         REXML::Document.new(f)
       end
-      assert_equal(utf16.encoding, "UTF-16")
+      assert_equal("UTF-16", utf16.encoding)
       assert( utf16[0].kind_of?(REXML::XMLDecl))
     end
+
+    def test_default_internal_with_utf16
+      orig_internal = ::Encoding.default_internal
+      ::Encoding.default_internal = 'utf-8'
+      begin
+        test_ticket_110
+      ensure
+        ::Encoding.default_internal = orig_internal
+      end
+    end
   end
 end

[usa (Usaku NAKAMURA)]

• Related to Bug #10417: IO#set_encoding without int_enc doesn't keep current internal encoding added

[nobu (Nobuyoshi Nakada)]

• Description updated (diff)

-wで警告が出そうなので、EnvUtil.with_default_internalを使うかこれに相当するようにしたほうがいいんじゃないでしょうか。

[kou (Kouhei Sutou)]

• Status changed from Assigned to Closed
• % Done changed from 0 to 100

Applied in changeset r48109.

---

• lib/rexml/source.rb (REXML::IOSource#encoding_updated): Fix a
  bug that can't parse XML correctly when
  Encoding.default_internal is different with XML
  encoding. REXML::Source converts XML encoding on read. So IO
  should not convert XML encoding.
  Based on patch by NAKAMURA Usaku.
  [ruby-dev:48686] [Bug #10418]

• test/rexml/test_encoding.rb
  (REXMLTests::EncodingTester#test_parse_utf16_with_utf8_default_internal):
  Add the for the above case.

[kou (Kouhei Sutou)]

EnvUtil.with_default_internalを使ったりしながら取り込みました！