Implementation difference. The command is always encoded into UTF-8, not the the encoding of the OS shell.
Added:
> >
RUN with passing any field of any type causes problems (C-asserts, or * glibc detected * free(): invalid next size (normal): 0x09d4ae20 ***g) if you run it twice. Run the RUN subroutine in TST to see the errors.
PASSing a unicode field via RUN doesn't seem to work, ie, echo $1 >some.file where $1 was a unicode field containing "This is Uni data\u03a3" does not give the expected result in the file (should contain a Σ symbol).
-- PeteBrower - 2011-03-17
\ No newline at end of file