Today, Joerg-Cyril Hoehle Joerg-Cyril.Hoehle@t-systems.com wrote:

I'm adding a few test cases to iterate-test.lisp. I wonder about the interaction of in-vector and by when moving backwards.

ITER> (iter (for i index-of-vector #(0 1 2 3 4) downto 0 by 2) (collect i)) (3 1)

I'd have expected (4 2 0) here. What do you think?

Sounds sensible. The analogy to (for x downfrom ...) sounds sensible. There is another interpretation, though:

(index-of-vector ... downto x) should go through the vector in the same way that the downto-less form would, in reverse order.

So these forms could also return: (iter (for i index-of-vector #(0 1 2 3 4) downto 0 by 2) (collect i)) ; (4 2 0)

(iter (for i index-of-vector #(0 1 2 3 4) downto 0 by 3) (collect i)) ; (3 0)

I'm undecided as to which is preferable. Your interpretation has the advantage that it's more in line with the behavior of the counting FOR driver, so I'm tending towards that.

Rationale: the first index/element it the rightmost, then walk by N.

The current behaviour looks to me like an unintended side effect of the implementation, moving down from (length array)=5 by 2 to yield 3,1.

Yeah, I can't imagine an interpretation where the current result makes much sense, either. (-:

Cheers,