What do you mean by "A(n) Unicode file"? UTF-8, UTF-16, UCS-2, or what? The <RS> construct, without further telling it otherwise, will work for UTF-8 but not the others. In UTF-8 you don't need "binmode" either.

If the file is using some other encoding, you need to set the input record separator to the proper byte sequence, and also use binmode.

In Perl 5.8, there is built-in support for reading files in other encodings. You can use the extended open syntax to specify, and all should work fine without further intervention.

—John