Ensure meta_encoding checks meta charset tag (closes #919)

sparklemotion · Jun 14, 2013 · fb000c5 · fb000c5
1 parent 14b7dc2
commit fb000c5
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 1 deletion.
diff --git a/lib/nokogiri/html/document.rb b/lib/nokogiri/html/document.rb
@@ -6,8 +6,12 @@ class Document < Nokogiri::XML::Document
       # then nil is returned.
       def meta_encoding
         meta = meta_content_type and
-          match = /charset\s*=\s*([\w-]+)/i.match(meta['content']) and
+          match = /charset\s*=\s*([\w-]+)/i.match(meta['content'])
+        if match
           match[1]
+        else
+          cs = at('meta[@charset]') and cs[:charset]
+        end
       end
 
       ###

diff --git a/test/html/test_document.rb b/test/html/test_document.rb
@@ -166,6 +166,20 @@ def test_meta_encoding_handles_malformed_content_charset
         assert_nil doc.meta_encoding
       end
 
+      def test_meta_encoding_checks_charset
+        doc = Nokogiri::HTML(<<-eohtml)
+<html>
+  <head>
+    <meta charset="UTF-8">
+  </head>
+  <body>
+    foo
+  </body>
+</html>
+        eohtml
+        assert_equal 'UTF-8', doc.meta_encoding
+      end
+
       def test_meta_encoding=
         @html.meta_encoding = 'EUC-JP'
         assert_equal 'EUC-JP', @html.meta_encoding