> Date: Thu, 19 Dec 2002 10:42:27 -0800
> From: Michael Lazzaro <[EMAIL PROTECTED]>
>
> Honestly, I still don't see what's so evil about R2L as:
>
> @out = sort given map {...} given grep {...} given @a;
>
> It seems to solve all the issues with not needing parens, not requiring
> {...} to magically eat a trailing comma but-only-sometimes, not needing
> to declare C<grep>, etc, in more than one place. And it's arguably
> more descriptive than the older (albeit shorter) syntax. And a *lot*
> more generically useful for things other than lists/arrays.
As it stands, there's a couple of problems with it. First, IIRC you
can't chain post-modifier things. Unless the likely has occurred that
Larry changed his mind from 8 months ago :).
Another is that C<sort> and company I<should> be methods of Array
(inherited or not), and if they're only in one place, then your
example would have to be:
@out = .sort given .map({...}) given .grep({...}) given @a;
The last problem is that R2L is not the problem. It is supported by
the basic syntax of the language:
@out = sort map {...} grep {...} @a;
Or some punctuational variant thereof.
If that doesn't work I can honestly say that Perl 6 is not Perl
(provided C<sort> and company are declared as forwarder calls, which I
don't see a problem with).
L2R is the big one that I'm hoping for. The pipe syntax is such a
powerful concept from Unix, I'm surprised Perl hasn't stolen it
before.
> $obj.foo(...);
> foo ... given $obj; # identical
I don't see how that works. My interpretation:
$obj.foo(...);
.foo(...) given $obj; # identical
Your way works for topic-defaulting calls, as in:
sub foo(my $arg //= $topic) is given($topic) {...}
foo $obj;
foo given $obj; #identical
Perhaps I'm missing something?
Luke
