Malcolm Wallace wrote: > I have been considering moving the lexer to use > ByteString instead of String, which would neatly solve that problem too. Doesn't the lexer come only after decoding? Then you have Unicode. Does ByteString still help? -Yitz