The new branch also enables testing of Encode.
I've now audited the remaining tests skipped under EBCDIC, and as a
result, some I've tweaked so they should work on EBCDIC, some I'm not
sure why they were skipped, so am enabling them to be tested to see why,
some should be skipped on EBCDIC for various good reasons, and some are
just too painful to port and do it maintainably. These are generally
UTF-8 malformations that don't port well to UTF-EBCDIC.
