Non-alphanumeric bogus character reference isn't a parse error.

nolanw · gsnedders · commit b872d927f15f · 2013-06-26T23:22:05.000+01:00
diff --git a/tokenizer/test4.test b/tokenizer/test4.test
@@ -46,7 +46,7 @@
 
 {"description":"Text after bogus character reference",
 "input":"<z z='&xlink_xmlns;'>bar<z>",
-"output":["ParseError",["StartTag","z",{"z":"&xlink_xmlns;"}],["Character","bar"],["StartTag","z",{}]]},
+"output":[["StartTag","z",{"z":"&xlink_xmlns;"}],["Character","bar"],["StartTag","z",{}]]},
 
 {"description":"Text after hex character reference",
 "input":"<z z='&#x0020; foo'>bar<z>",
