Document attribute charset and its aliases to return UTF-8

LayoutTests/dom/xhtml/level3/core/documentgetinputencoding02.js

Index: LayoutTests/dom/xhtml/level3/core/documentgetinputencoding02.js
diff --git a/LayoutTests/dom/xhtml/level3/core/documentgetinputencoding02.js b/LayoutTests/dom/xhtml/level3/core/documentgetinputencoding02.js
index 7e31b6af714b86b4b975608ba6f2d6e32ba02f92..bcd46020b0fe653da02505ed1731ebf7e445532e 100644
--- a/LayoutTests/dom/xhtml/level3/core/documentgetinputencoding02.js
+++ b/LayoutTests/dom/xhtml/level3/core/documentgetinputencoding02.js
@@ -104,7 +104,7 @@ docElem = doc.documentElement;
       newDoc = domImpl.createDocument(rootNS,rootName,nullDocType);
       encodingName = newDoc.inputEncoding;
 
-      assertNull("documentgetinputencoding02",encodingName);
+      assertEquals("documentgetinputencoding02", "UTF-8".toLowerCase(), encodingName.toLowerCase());

[dominicc (has gone to gerrit), 2015/06/22 01:23:31]

The spec your change mentions links to this to explain what the name of UTF-8
is:

https://encoding.spec.whatwg.org/#names-and-labels

And what you're asserting here accepts invalid names and doesn't accept all of
the valid ones (you're missing "unicode-1-1-utf-8".) Write an accurate
assertion. Unless there's a spec I'm missing... link?

[Habib Virji, 2015/06/22 14:07:30]

The spec link I mentioned was to emphasise it does not return null but rather
utf-8. Both Gecko and Blink set the value as UTF-8 and not utf-8, utf8 or
unicode-1-1-utf-8. I have updated the test to reflect this.

[philipj_slow, 2015/06/23 08:32:56]

"utf8" and "unicode-1-1-utf-8" don't matter here, those are labels that map to
the internal enum or atomic string representing UTF-8, while this should be
testing how that enum or atomic string is represented when converted back to a
string.

 
 }