NAME
OO::InsideOut - Minimal support for Inside-Out Classes
VERSION
0.03
SYNOPSIS
package My::Name;
use OO::InsideOut qw(id register);
register \my( %Name, %Surname );
sub new {
my $class = shift;
return bless \(my $o), ref $class || $class;
}
sub name {
my $id = id( shift );
scalar @_
and $Name{ $id } = shift;
return $Name{ $id };
}
sub surname {
my $id = id( shift );
scalar @_
and $Surname{ $id } = shift;
return $Surname{ $id };
}
...
EXPORT
Nothing by default but every function, in FUNCTIONS, can be exported on demand.
DESCRIPTION
NOTE: If you're developing for perl 5.10 or later, please consider using Hash::Util::FieldHash instead.
OO::InsideOut provides minimal support for Inside-Out Classes for perl 5.8 or later. By minimal, understand;
No special methods or attributtes;
Don't use source filters or any kind of metalanguage;
No need for a special constructor;
No need to register objects;
No serialization hooks (like Storable, Dumper, etc);
It provides:
Automatic object registration;
Automatic object destruction;
Thread support (but not shared);
Foreign inheritance;
mod_perl compatibility
FUNCTIONS
id
id( $object );
Uses Scalar::Util::refaddr to return the reference address of $object.
register
register( @hashrefs );
Register the given hashrefs for proper cleanup.
Returns an HASH ref with registered objects in the CLASS. See CAVEATS.
Dumper
Dumper( $object );
If available, uses Data::Dumper::Dumper to dump the object's data.
WARNING: May be removed in the future!!!
HOW IT WORKS
When registering hashes, and only then, OO::InsideOut will:
Wrap any new() method*, in the inheritance tree, with the ability to register objects;
Wrap any DESTROY() method*, in the inheritance tree, with the ablity to cleanup the object's data;
If no DESTROY() method was found, it provides one in the firs package of the inheritance tree;
* This is done only once per package.
PERFORMANCE
Every Inside-Out technique, using an id to identify the object, will be slower than the classic OO approach: it's just the way it is.
Consider:
sub name {
my $self = shift;
scalar @_
&& $Name{ id( $self ) } = shift;
return $Name{ id( $self ) );
}
In this example, the code is calling the id twice, causing uncessary overload. If you are going to use id more than once, in the same scope, consider saving it in an variable earlier:
sub name {
my $id = id( shift );
scalar @_
&& $Name{ $id } = shift;
return $Name{ $id };
}
MIGRATING TO Hash::Util::FieldHash
Bare in mind that, besides the obvious diferences between the two modules, in Hash::Util::FieldHash, the cleanup process is triggered before calling DESTROY(). In OO::Insideout, this only happens after any DESTROY() defined in the package.
DIAGNOSTICS
- must provide, at least, one hash ref!
-
Besides the obvious reason, this migth happen while using
my
with a list with only one item:register \my( %Field ) #WRONG register \my %Field #RIGTH
CAVEATS
register(), on request, will return an HASH ref with all the objects registered in the CLASS.
If, for any reason, you need to copy/grep this HASH ref, make sure to weaken every entry again. See Scalar::Util::weaken for more detail on this subject.
AUTHOR
André "Rivotti" Casimiro, <rivotti at cpan.org>
BUGS
Please report any bugs or feature requests through the issue tracker at https://github.com/ARivottiC/OO-InsideOut/issues.
SUPPORT
You can find documentation for this module with the perldoc command.
perldoc OO::InsideOut
You can also look for information at:
GitHub
AnnoCPAN: Annotated CPAN documentation
CPAN Ratings
Search CPAN
ACKNOWLEDGEMENTS
SEE ALSO
Alter, Class::InsideOut, Class::Std, Hash::Util::FieldHash, Object::InsideOut.
LICENSE AND COPYRIGHT
Copyright 2013 André Rivotti Casimiro.
This program is free software; you can redistribute it and/or modify it under the terms of either: the GNU General Public License as published by the Free Software Foundation; or the Artistic License.
See http://dev.perl.org/licenses/ for more information.