Michael Fowler wrote:
>
> This RFC makes no mention of what happens to the following constructs:
>
> use Foo;
> require Foo;
I don't mind a very few special cases. I would imagine that these might
have to remain special:
use Foo;
require Foo;
import Foo;
package Foo;
But notice the disconnect here:
package Foo;
use base 'Bar';
Seems almost that:
package 'Foo';
use base 'Bar';
Would be more consistent. The only real problem with this approach is
require() which, as Schwern has to keep reminding me, can work on either
files or packages:
require Foo; # package Foo
require 'Foo'; # file Foo
But this already seems fragile to me. Why not have a separate keyword
that always assumes it's a file? Decreased bloat; require could always
be UNIVERSAL::require then:
require 'Foo'; # package Foo
include 'Foo'; # file Foo
This would get rid of having to eval stuff for variable-based package
names as well.
> %foo = (bar => "baz");
This would remain as-is, since => auto-quotes the LH operand.
-Nate
