Possible encoding problem in HTTP headers when showing webpages

In ARCArchiveAccess HTTP headers are parsed and put in HttpResponse in the
process converting between a bytestream and a String with default encoding.
This might go wrong if the actual HTTP header contains non-ascii characters in a
different encoding than our default encoding. It is expected but not checked
that current HTTP headers probably don't contain non-ascii characters.
It is unknown how you can actually detect the HTTP header encoding. There is a
field for body-encoding but a quick search didn't reveal how you find out
about the header-encoding.
NOTE: This bug is originally from Bugzilla bug_id=760.
This bug was previously assigned to Unassigned.