This same behavior exists in 5.004. Notice a similar "problem" with chop( @a[0..2] ).

It certainly seems like a bug but, I have to say, I can't get too worked up about it as being a serious bug. But a bug report is in order.

        - tye (but my friends call me "Tye")