>v) @x->[2] and scalar(@x)->[2] shouldn't be lumped together. As > discussed on p5p long ago (pre 5.004 ?), what @x->[2] does is > deliberate, if unfortunate. Whereas having scalar(@x) > evaluate to anything other than the length of @x is a bug. So, you're saying that it is desirable that @x act one way in an implicit scalar context and another way in an explicit one? --tomThread Previous | Thread Next