Early feedback on the new ByteToCharCOMPOUND_TEXT and CharToByteCOMPOUND_TEXT
converters indicates that they should be expanded to include support for
additional, non-ISO-2022 encodings. Adding support for these encodings is
trivial, except that it expands the list of legacy converters which must remain
in the JRE until the x-compound-text converters are switched over to the
java.nio API.

The encodings koi8-r and cp1251 were specifically mentioned by Valeriy Ushakov.
Apparently, use of ISO-8859-5 for Cyrillic text is uncommon. The full list of
supported encodings should be reevaluated for merlin-beta2.