NAME

File::Find::Rule::Extending - the mini-guide to extending File::Find::Rule

SYNOPSIS

 package File::Find::Rule::Random;
 use strict;
 
 # take useful things from File::Find::Rule
 use base 'File::Find::Rule';

 # and force our crack into the main namespace
 sub File::Find::Rule::random () {
     my $self = shift()->_force_object;
     $self->exec( sub { rand > 0.5 } );
 }
 
 1;

DESCRIPTION

File::Find::Rule went down so well with the buying public that every wanted to add extra features. With the 0.07 release this became a possibility, using the following conventions.

Declare your package

 package File::Find::Rule::Random;
 use strict;

Inherit methods from File::Find::Rule

 # take useful things from File::Find::Rule
 use base 'File::Find::Rule';
 
=head3 Force your madness into the main package

 # and force our crack into the main namespace
 sub File::Find::Rule::random () {
     my $self = shift()->_force_object;
     $self->exec( sub { rand > 0.5 } );
 }

Yes, we're being very cavalier here and defining things into the main File::Find::Rule namespace. This is due to lack of imaginiation on my part - I simply can't find a way for the functional and oo interface to work without doing this or some kind of inheritance, and inheritance stops you using two File::Find::Rule::Foo modules together.

For this reason try and pick distinct names for your extensions. If this becomes a problem then I may institute a semi-official registry of taken names.

Taking no arguments.

Note the null prototype on random. This is a cheat for the procedural interface to know that your sub takes no arguments, and so allows this to happen:

 find( random => in => '.' );

If you hadn't declared random with a null prototype it would have consumed in as a parameter to it, then got all confused as it doesn't know about a '.' rule.

AUTHOR

Richard Clamp <richardc@unixbeard.net>

COPYRIGHT

This module is free software; you can redistribute it and/or modify it under the same terms as Perl itself.