switch html5lib to new pkg layout

test/tokenizer_test.dart

 #library('tokenizer_test');
 
 // Note: mirrors used to match the getattr usage in the original test
 #import('dart:io');
 #import('dart:json');
 #import('dart:mirrors');
 #import('package:unittest/unittest.dart');
 #import('package:unittest/vm_config.dart');
-#import('../lib/char_encodings.dart');
-#import('../lib/constants.dart', prefix: 'constants');
-#import('../lib/token.dart');
-#import('../lib/utils.dart');
-#import('../tokenizer.dart');
+#import('package:html5lib/src/char_encodings.dart');
+#import('package:html5lib/src/constants.dart', prefix: 'constants');
+#import('package:html5lib/src/token.dart');
+#import('package:html5lib/src/utils.dart');
+#import('package:html5lib/tokenizer.dart');
 #import('support.dart');
 
 /**
  * This is like [JSON.parse], but it fixes unicode surrogate pairs in the JSON.
  *
  * Without this, the test "expects" incorrect results from the tokenizer.
  * Note: Python's json module decodes these correctly, so this might point at
  * a bug in Dart's [JSON.parse].
  */
 jsonParseUnicode(String input) => jsonFixSurrogatePairs(JSON.parse(input));
@@ skipping 271 matching lines @@
             test(testInfo["description"], () {
               testInfo["initialState"] = camelCase(initialState);
               runTokenizerTest(testInfo);
             });
           }
         }
       });
     }
   });
 }