Multilingual input

Hi,

I am trying to build a multilingual input decoder for Indic languages. The
host locale is 'en' but keystrokes may contain Indic language codes in UTF-8
encoding. However, 'en' language is under Latin1Environment right now and
assumes MacRoman for codes > 127. I suppose I could inject a global preference
to resolve the conflict but only if it is absolutely necessary.

Instead, I wanted to modify the Latin1 inputInterpreter to use UTF-8 decoder
from UnixUTF8JPInputInterpreter on Unix when code > 127.

Are there Unix VMs that feed MacRoman encodings into Smalltalk in evt third
but keep 0 in evt sixth for Latin1 locales?

Thanks .. Subbu