package ProcInfo;

use Carp;
use strict;
use warnings;


our $VERSION = '0.01';

sub new 
{
  my $class = shift;
  my $self = {};
  bless $self, $class;
  $self->_init();
  return $self;
}

sub _init
{
  my $self = shift;
  my $osModule = "ProcInfo_$^O";

  # Load the subclass if it isn't already loaded.
  eval "require $osModule";
  if($@)
  {
    croak "Could not load OS ProcInfo Module:$osModule. [$@]";
  }
  
  # rebless self using subclass.
  bless $self, $osModule;
  $self->_init(@_);
  
  return $self;
}

sub method_1
{
  return 1;
}

1;
[download]

Some question I have about the base class implementation:

• If I know that I'm going to rebless the object to a subclass. Is there any advantage to blessing it in the first place in the new?
• Is the rebless in the _init the right way to recast self to another class?

The skeleton for each sublcass would looking like this and it would be stored in ProcInfo_linux.pm:

package ProcInfo_linux;

use strict;
use warnings;
use base qw|ProcInfo|;

our $VERSION = '0.01';

sub _init
{
  my $self = shift;
  return $self;
}

sub method_1
{
  return 1;
}

1;
[download]

Finally, here is my test script (ProcInfo.t):

use Test::More tests => 3;
BEGIN { use_ok('ProcInfo') };


my $pInfo = new ProcInfo;
is(ref($pInfo), "ProcInfo_$^O", "Factory new Works");
ok($pInfo->method_1());
[download]