Yes, sorry about that BazB. The pseudocode was only illustrative :).

One thing with your solution though. I've seen merlyn at least, and possibly others condemn the practice of using

sub new {
    my $proto = shift;
    my $class = ref( $proto ) || $proto;
    my $self = ....;
    ...
    return bless $self, $class;
}
[download]

as "cargo-cultish" behaviour. This is a somewhat condensed version of your solution.

The problem apparently being (if I understand the argument correctly), that it makes no distinction between calling Foo->new(); to create a new instance and calling $instance->new(); to make a clone of an existing instance. I'm not sure that this affects you in this case, but merlyn is much wiser in these things than I, so you probably ought to read his advice for yourself.

I found one reference of his at ref($proto) - just say no! which goes way back. I know that there have been many more instances in the more recent past. If your lucky, he'll get over the fact that I'm involved and give you the gen on this himself.

Update: I found another, slightly more recent discussion, including an alternative method of achieving what you need to achieve (by putting the onus on the caller rather than in the new() call) here, and a better explanation of the issue by dragonchild here. In fact, the whole thread is well worth a read as most of the big hitters around here step in and argue the case, which is a great help in letting you make up your own mind on the issues.

It helped me a form an opinion anyway:)