Outer product considered useful

[Luke Palmer]

I found myself writing a perl script today in which I did (I'll
perl6-ize it for sake of discussion):

    for 98,99 -> $i {
        for 0..255 -> $j {
            # testing IP addresses with $i.$j
        }
    }

I was thinking about what would happen if I allowed the user to input
those ranges, like:

    67.173.[98,99].[0..255]

After appropriate data massagng, I'd have to do something like:

    for @(@ranges[1]) -> $i {
        for @(@ranges[2]) -> $j {
            for @(@ranges[3]) -> $k {
                for @(@ranges[4]) -> $l {
                    # ...
                }
            }
        }
    }

Using a permutations module I could make that shorter, but I figure that
since we're already providing C<zip> to make looping easier, why not
provide C<outer> (perhaps spelled Â)?  The outer function would provide
a way to I<dynamically> nest loops.   

    for outer([EMAIL PROTECTED]) -> [ $i, $j, $k, $l ] 
    {
        # ...
    }

Or better yet:

    for outer([EMAIL PROTECTED]) -> @cp {
        my $ip = @cp.join('.');
    }

And the original loop could be written:

    for 98,99 Â 0..255 -> [$i, $j] {
        # ...
    }

Supposing  had sufficiently low precedence.  And supposing  were used
at all, something I'm not particuarly attached to happening.

I believe it could be programmed lazily.  Like this:

    sub _outer_coro(*$first is context(Scalar), 
                    [EMAIL PROTECTED]  is context(Scalar))
        is coroutine
    {
        if @rest {
            _outer_coro [EMAIL PROTECTED];
        }
        else {
            yield $first;
        }
    }

    sub outer([EMAIL PROTECTED] is context(Scalar))
    {
        <_outer_coro([EMAIL PROTECTED])>
    }

Luke