anchors and aliases have an encoding

2 files changed, 84 insertions, 14 deletions

diff --git a/ext/psych/parser.c b/ext/psych/parser.c
index e4a6e98..76ca4dd 100644
--- a/ext/psych/parser.c
+++ b/ext/psych/parser.c
@@ -132,11 +132,15 @@ static VALUE parse(VALUE self, VALUE yaml)
         );
         break;
       case YAML_ALIAS_EVENT:
-        rb_funcall(handler, id_alias, 1,
-          event.data.alias.anchor ?
-          rb_str_new2((const char *)event.data.alias.anchor) :
-          Qnil
-        );
+        {
+          VALUE alias = Qnil;
+          if(event.data.alias.anchor) {
+            alias = rb_str_new2((const char *)event.data.alias.anchor);
+            rb_enc_associate_index(alias, encoding);
+          }
+
+          rb_funcall(handler, id_alias, 1, alias);
+        }
         break;
       case YAML_SCALAR_EVENT:
         {
@@ -147,13 +151,17 @@ static VALUE parse(VALUE self, VALUE yaml)
 
           rb_enc_associate_index(val, encoding);
 
-          VALUE anchor = event.data.scalar.anchor ?
-            rb_str_new2((const char *)event.data.scalar.anchor) :
-            Qnil;
+          VALUE anchor = Qnil;
+          if(event.data.scalar.anchor) {
+            anchor = rb_str_new2((const char *)event.data.scalar.anchor);
+            rb_enc_associate_index(anchor, encoding);
+          }
 
-          VALUE tag = event.data.scalar.tag ?
-            rb_str_new2((const char *)event.data.scalar.tag) :
-            Qnil;
+          VALUE tag = Qnil;
+          if(event.data.scalar.tag) {
+            tag = rb_str_new2((const char *)event.data.scalar.tag);
+            rb_enc_associate_index(tag, encoding);
+          }
 
           VALUE plain_implicit =
             event.data.scalar.plain_implicit == 0 ? Qfalse : Qtrue;
@@ -173,9 +181,11 @@ static VALUE parse(VALUE self, VALUE yaml)
             rb_str_new2((const char *)event.data.sequence_start.anchor) :
             Qnil;
 
-          VALUE tag = event.data.sequence_start.tag ?
-            rb_str_new2((const char *)event.data.sequence_start.tag) :
-            Qnil;
+          VALUE tag = Qnil;
+          if(event.data.sequence_start.tag) {
+            tag = rb_str_new2((const char *)event.data.sequence_start.tag);
+            rb_enc_associate_index(tag, encoding);
+          }
 
           VALUE implicit =
             event.data.sequence_start.implicit == 0 ? Qfalse : Qtrue;
diff --git a/test/psych/test_encoding.rb b/test/psych/test_encoding.rb
new file mode 100644
index 0000000..d97856f
--- /dev/null
+++ b/test/psych/test_encoding.rb
@@ -0,0 +1,60 @@
+# -*- coding: utf-8 -*-
+
+require 'minitest/autorun'
+require 'psych'
+
+module Psych
+  class TestEncoding < MiniTest::Unit::TestCase
+    class EncodingCatcher < Handler
+      attr_reader :strings
+      def initialize
+        @strings = []
+      end
+
+      (Handler.instance_methods(true) -
+       Object.instance_methods).each do |m|
+        class_eval %{
+          def #{m} *args
+            @strings += args.flatten.find_all { |a|
+              String === a
+            }
+          end
+        }
+      end
+    end
+
+    def setup
+      super
+      @handler = EncodingCatcher.new
+      @parser  = Psych::Parser.new @handler
+      @utf8    = Encoding.find('UTF-8')
+    end
+
+    def test_scalar
+      @parser.parse("--- a")
+      assert_encodings @utf8, @handler.strings
+    end
+
+    def test_alias
+      @parser.parse(<<-eoyml)
+%YAML 1.1
+---
+!!seq [
+  !!str "Without properties",
+  &A !!str "Anchored",
+  !!str "Tagged",
+  *A,
+  !!str "",
+]
+      eoyml
+      assert_encodings @utf8, @handler.strings
+    end
+
+    private
+    def assert_encodings encoding, strings
+      strings.each do |str|
+        assert_equal encoding, str.encoding, str
+      end
+    end
+  end
+end