I noticed in SOAP::Lite 0.70_2 you fixed the UTF8 and base64 bug where
base64_encode() was handed UTF8 data (CPAN Bug# 30271).
We're using SOAP::Lite 0.60_1 for compatibility reasons so I'm looking at
the best way to back-port this fix.
1) On deserialization, does something else detect the bytes as UTF8 and
mark them as UTF8 strings in Perl? The base64 serialization routines have
nothing special in them to do this.
2) Is (1) why you serialize with _utf8_off()/pack() rather than some kind
of Encode::encode/decode() pairing?