Re: Re: Re: Re: method of ID'ing

[Juerd]

how about adding require Tie::Scalar

There's not much point in having a base class if you override all methods :) Well, I don't use new() and DESTROY(), but those are rather optional...

and throwing this in the code catacombs? this is a nice, simple answer to retrieving the current time.

It doesn't belong there. This use should not be encouraged, just use time instead. This was a one-minute hack, typed in my browser. If you think it belongs there, go ahead, there is no copyright at all :)

While you're at it, create these as well:

package Tie::Print;
require Tie::Scalar; our @ISA = ('Tie::StdScalar');
sub STORE { print pop }

package Tie::Localtime::Scalar;
require Tie::Scalar; our @ISA = ('Tie::StdScalar');
sub FETCH { return scalar localtime }

etcetera

Have a look at perlfunc for more ideas. Or just use a one-fits all tie:

package Tie::Scalar::Miscellaneous;
use Carp;
use strict;

sub TIESCALAR {
    my ($class, %functions) = @_;
    $_ ||= q(croak 'Not implemented') for @functions{qw/store fetch/};
    $_ = eval "sub { $_ ( \@_ ) }" for @functions{qw/store fetch/};
    bless \%functions, $class;
}

sub STORE {
    shift->{store}->(pop)
}

sub FETCH {
    shift->{fetch}->();
}

Which can be used like:

tie my $printer, 'Tie::Scalar::Miscellaneous', store => 'print';
$printer = "Hello, world!\n";

No, I'm not being serious.

- Yes, I reinvent wheels.
- Spam: Visit eurotraQ.

[particle]

package Tie::Scalar::Miscellaneous;

ha! i love it. who needs functions when everything can be a variable assignment?

by the way, i agree that time should be used instead. i thought perhaps this would be a good fit for something like gav^ or tachyon's solutions (posted above), although it's probably easier to maintain as a function rather than a tied scalar.

~Particle ;Þ