On Thu, Feb 28, 2002 at 03:28:29PM -0800, Austin Hastings wrote:
>
> --- Garrett Goebel <[EMAIL PROTECTED]> wrote:
> >
> > Hmm... Out of curiosity what kind of user-extensible topicalizer
> > aware constructs would you make?
>
> Remember Larry's comment that the -> operator is a kind of "parameter
> binding", and not something that's exclusive to given/when.
>
> The example: my $closure = -> $a { ... };
>
> The point made was that sub($a) { ... }
> would be different because it paid more attention to (I think) the
> stack frame-y bits.
>
> I'm (optimistically) inferring that means that we'll be able to bind a
> non-sub closure in the current context.
>
> maybe something like:
>
> my $closure = -> $a { ... };
>
> sub doit(&)
> {
> my $cl = shift;
> my @ary = ...;
>
> for @ary -> $cl;
> }
>
> (Larry denied this once, but I suspect it's too close to doable for
> Damian to let it slip by... )
It seems anti-dwim. Once people get used to the idea that "for @x -> $y"
means that $y holds the values of @x, I think it would just be annoying
to totally shift that meaning by having a $y in what looks like the same
position do something entirely different. It's overusing the operator.
> > Larry Wall in Apocalypse 4 writes:
> > > A when is the only defaulting construct that pays attention
> > > to the current topicalizer regardless of which variable it
> > > is associated with.
>
> Frankly, I'm not sure that -> and "topicalizer" are the same. I get the
> feeling that -> is really "binding operator" and "topicalizer" is
> 'given' (with the caveat, of course, that 'for' has an implicit
> 'given').
Yes, C<given> and C<for> are topicalizers, -> is not.
> I say this because some of the other examples of using -> don't have
> any topicality to them, or they can be read to imply SO MUCH topicality
> that the notion becomes worthless (think $_ in perl3/4).
>
> If we have: my $closure = -> $a { ... };
>
> are we really to take $a as not only an arg to $closure but the default
> topic within $a?
Do you mean "the default topic within the closure"? Yes, but (if this
worked) it wouldn't be -> that did it.
> Larry Wall in Apocalypse 4 writes:
> > Bare closures topicalize their first argument...
It's a behavior of a closure, with or without the addition of aliasing.
Allison
