Refactored the request charset decoding process. It now only decodes
request lines and POST form responses, not everything and then some
that happens to be in the request object.
The charset for incoming requests now defaults to utf-8 instead of
iso-8859-1. There is a silent fallback to iso-8859-1 if utf-8 decode
fails, and magic_roxen_automatic_charset_variable takes precedence if