Fix bug #706: 2nd try. Implement the charset rules as follows:

SRU SOAP / POST: If charset in content-type is specified, use it . If
omitted, do not assume a particular charset (And &#x.. is in use).
SRU GET: If charset in content-type is specified, use it. If omitted, assume
UTF-8. Note that in most cases there will not even be a content-type for
SRU GET.