* ext/psych/emitter.c: output strings are automatically transcoded

* test/psych/test_emitter.rb: supporting tests
* test/psych/test_encoding.rb: more supporting tests

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@27910 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

ChangeLog

@@ -1,3 +1,7 @@
+Thu May 20 02:12:20 2010  Aaron Patterson <aaron@tenderlovemaking.com>
+
+	* ext/psych/emitter.c: output strings are automatically transcoded
+
 Wed May 19 23:36:57 2010  Hidetoshi NAGAI  <nagai@ai.kyutech.ac.jp>
 
 	* ext/tk/extconf.rb: [ruby-dev:41334] [Bug #3307] invalid result

ext/psych/emitter.c

@@ -121,6 +121,9 @@ static VALUE start_document(VALUE self, VALUE version, VALUE tags, VALUE imp)
 
     if(RTEST(tags)) {
 	int i = 0;
+#ifdef HAVE_RUBY_ENCODING_H
+	rb_encoding * encoding = rb_utf8_encoding();
+#endif
 
 	Check_Type(tags, T_ARRAY);
 
@@ -129,15 +132,24 @@ static VALUE start_document(VALUE self, VALUE version, VALUE tags, VALUE imp)
 
 	for(i = 0; i < RARRAY_LEN(tags); i++) {
 	    VALUE tuple = RARRAY_PTR(tags)[i];
+	    VALUE name;
+	    VALUE value;
+
 	    Check_Type(tuple, T_ARRAY);
 
 	    if(RARRAY_LEN(tuple) < 2) {
 		xfree(head);
 		rb_raise(rb_eRuntimeError, "tag tuple must be of length 2");
 	    }
+	    name  = RARRAY_PTR(tuple)[0];
+	    value = RARRAY_PTR(tuple)[1];
+#ifdef HAVE_RUBY_ENCODING_H
+	    name = rb_str_export_to_enc(name, encoding);
+	    value = rb_str_export_to_enc(value, encoding);
+#endif
 
-	    tail->handle = (yaml_char_t *)StringValuePtr(RARRAY_PTR(tuple)[0]);
+	    tail->handle = (yaml_char_t *)StringValuePtr(name);
-	    tail->prefix = (yaml_char_t *)StringValuePtr(RARRAY_PTR(tuple)[1]);
+	    tail->prefix = (yaml_char_t *)StringValuePtr(value);
 
 	    tail++;
 	}
@@ -199,6 +211,22 @@ static VALUE scalar(
 
     Check_Type(value, T_STRING);
 
+#ifdef HAVE_RUBY_ENCODING_H
+    rb_encoding * encoding = rb_utf8_encoding();
+
+    value = rb_str_export_to_enc(value, encoding);
+
+    if(!NIL_P(anchor)) {
+	Check_Type(anchor, T_STRING);
+	anchor = rb_str_export_to_enc(anchor, encoding);
+    }
+
+    if(!NIL_P(tag)) {
+	Check_Type(tag, T_STRING);
+	tag = rb_str_export_to_enc(tag, encoding);
+    }
+#endif
+
     yaml_scalar_event_initialize(
 	    &event,
 	    (yaml_char_t *)(NIL_P(anchor) ? NULL : StringValuePtr(anchor)),
@@ -231,6 +259,21 @@ static VALUE start_sequence(
 	) {
     yaml_emitter_t * emitter;
     yaml_event_t event;
+
+#ifdef HAVE_RUBY_ENCODING_H
+    rb_encoding * encoding = rb_utf8_encoding();
+
+    if(!NIL_P(anchor)) {
+	Check_Type(anchor, T_STRING);
+	anchor = rb_str_export_to_enc(anchor, encoding);
+    }
+
+    if(!NIL_P(tag)) {
+	Check_Type(tag, T_STRING);
+	tag = rb_str_export_to_enc(tag, encoding);
+    }
+#endif
+
     Data_Get_Struct(self, yaml_emitter_t, emitter);
 
     yaml_sequence_start_event_initialize(
@@ -283,6 +326,20 @@ static VALUE start_mapping(
     yaml_event_t event;
     Data_Get_Struct(self, yaml_emitter_t, emitter);
 
+#ifdef HAVE_RUBY_ENCODING_H
+    rb_encoding * encoding = rb_utf8_encoding();
+
+    if(!NIL_P(anchor)) {
+	Check_Type(anchor, T_STRING);
+	anchor = rb_str_export_to_enc(anchor, encoding);
+    }
+
+    if(!NIL_P(tag)) {
+	Check_Type(tag, T_STRING);
+	tag = rb_str_export_to_enc(tag, encoding);
+    }
+#endif
+
     yaml_mapping_start_event_initialize(
 	    &event,
 	    (yaml_char_t *)(NIL_P(anchor) ? NULL : StringValuePtr(anchor)),
@@ -327,6 +384,13 @@ static VALUE alias(VALUE self, VALUE anchor)
     yaml_event_t event;
     Data_Get_Struct(self, yaml_emitter_t, emitter);
 
+#ifdef HAVE_RUBY_ENCODING_H
+    if(!NIL_P(anchor)) {
+	Check_Type(anchor, T_STRING);
+	anchor = rb_str_export_to_enc(anchor, rb_utf8_encoding());
+    }
+#endif
+
     yaml_alias_event_initialize(
 	    &event,
 	    (yaml_char_t *)(NIL_P(anchor) ? NULL : StringValuePtr(anchor))

test/psych/test_emitter.rb

@@ -64,6 +64,7 @@ module Psych
         ['foo', Object.new, nil, false, true, 1],
         ['foo', nil, Object.new, false, true, 1],
         ['foo', nil, nil, false, true, :foo],
+        [nil, nil, nil, false, true, 1],
       ].each do |args|
         assert_raises(TypeError) do
           @emitter.scalar(*args)

test/psych/test_encoding.rb

@@ -24,9 +24,81 @@ module Psych
 
     def setup
       super
+      @buffer  = StringIO.new
       @handler = EncodingCatcher.new
       @parser  = Psych::Parser.new @handler
       @utf8    = Encoding.find('UTF-8')
+      @emitter = Psych::Emitter.new @buffer
+    end
+
+    def test_emit_alias
+      @emitter.start_stream Psych::Parser::UTF8
+      @emitter.start_document [], [], true
+      e = assert_raises(RuntimeError) do
+        @emitter.alias 'ドラえもん'.encode('EUC-JP')
+      end
+      assert_match(/alias value/, e.message)
+    end
+
+    def test_start_mapping
+      foo = 'foo'
+      bar = 'バー'
+
+      @emitter.start_stream Psych::Parser::UTF8
+      @emitter.start_document [], [], true
+      @emitter.start_mapping(
+        foo.encode('Shift_JIS'),
+        bar.encode('UTF-16LE'),
+        false, Nodes::Sequence::ANY)
+      @emitter.end_mapping
+      @emitter.end_document false
+      @emitter.end_stream
+
+      @parser.parse @buffer.string
+      assert_encodings @utf8, @handler.strings
+      assert_equal [foo, bar], @handler.strings
+    end
+
+    def test_start_sequence
+      foo = 'foo'
+      bar = 'バー'
+
+      @emitter.start_stream Psych::Parser::UTF8
+      @emitter.start_document [], [], true
+      @emitter.start_sequence(
+        foo.encode('Shift_JIS'),
+        bar.encode('UTF-16LE'),
+        false, Nodes::Sequence::ANY)
+      @emitter.end_sequence
+      @emitter.end_document false
+      @emitter.end_stream
+
+      @parser.parse @buffer.string
+      assert_encodings @utf8, @handler.strings
+      assert_equal [foo, bar], @handler.strings
+    end
+
+    def test_doc_tag_encoding
+      key = '鍵'
+      @emitter.start_stream Psych::Parser::UTF8
+      @emitter.start_document(
+        [1, 1],
+        [['!'.encode('EUC-JP'), key.encode('EUC-JP')]],
+        true
+      )
+      @emitter.scalar 'foo', nil, nil, true, false, Nodes::Scalar::ANY
+      @emitter.end_document false
+      @emitter.end_stream
+
+      @parser.parse @buffer.string
+      assert_encodings @utf8, @handler.strings
+      assert_equal key, @handler.strings[1]
+    end
+
+    def test_emitter_encoding
+      str  = "壁に耳あり、障子に目あり"
+      thing = Psych.load Psych.dump str.encode('EUC-JP')
+      assert_equal str, thing
     end
 
     def test_default_internal