+=head2 _DecodeLOB C<ContentType>, C<ContentEncoding>, C<Content>
+
+Unpacks data stored in the database, which may be base64 or QP encoded
+because of our need to store binary and badly encoded data in columns
+marked as UTF-8. Databases such as PostgreSQL and Oracle care that you
+are feeding them invalid UTF-8 and will refuse the content. This
+function handles unpacking the encoded data.
+
+It returns textual data as a UTF-8 string which has been processed by Encode's
+PERLQQ filter which will replace the invalid bytes with \x{HH} so you can see
+the invalid byte but won't run into problems treating the data as UTF-8 later.
+
+This is similar to how we filter all data coming in via the web UI in
+RT::Interface::Web::DecodeARGS. This filter should only end up being
+applied to old data from less UTF-8-safe versions of RT.
+
+If the passed C<ContentType> includes a character set, that will be used
+to decode textual data; the default character set is UTF-8. This is
+necessary because while we attempt to store textual data as UTF-8, the
+definition of "textual" has migrated over time, and thus we may now need
+to attempt to decode data that was previously not trancoded on insertion.
+
+Important Note - This function expects an octet string and returns a
+character string for non-binary data.
+
+=cut
+