Utf8ConvertJob: the convert() method decode/encode logic is unfinished

The decode/encode logic in Utf8ConvertJob.convert() is unfinished.
1) We need to modify/add the Content-Encoding HTTP header field
2) Convert all HTTP headers to UTF-8 after modifying/adding Content-Encoding field.
3) Convert the body of the ARCRecord to UTF-8.
4) Concatenate the result of 2) with the result of 3)
5) convert to a single byte-array (byte[]).
NOTE: This bug is originally from Bugzilla bug_id=26.