class metaclasses can also be applied directly

[gitmo/Moose.git] / lib / Moose / Cookbook / Meta / Labeled_AttributeTrait.pod

package Moose::Cookbook::Meta::Labeled_AttributeTrait;

# ABSTRACT: Labels implemented via attribute traits

__END__


=pod

=head1 SYNOPSIS

  package MyApp::Meta::Attribute::Trait::Labeled;
  use Moose::Role;

  has label => (
      is        => 'rw',
      isa       => 'Str',
      predicate => 'has_label',
  );

  package Moose::Meta::Attribute::Custom::Trait::Labeled;
  sub register_implementation {'MyApp::Meta::Attribute::Trait::Labeled'}

  package MyApp::Website;
  use Moose;

  has url => (
      traits => [qw/Labeled/],
      is     => 'rw',
      isa    => 'Str',
      label  => "The site's URL",
  );

  has name => (
      is  => 'rw',
      isa => 'Str',
  );

  sub dump {
      my $self = shift;

      my $meta = $self->meta;

      my $dump = '';

      for my $attribute ( map { $meta->get_attribute($_) }
          sort $meta->get_attribute_list ) {

          if (   $attribute->does('MyApp::Meta::Attribute::Trait::Labeled')
              && $attribute->has_label ) {
              $dump .= $attribute->label;
          }
          else {
              $dump .= $attribute->name;
          }

          my $reader = $attribute->get_read_method;
          $dump .= ": " . $self->$reader . "\n";
      }

      return $dump;
  }

  package main;

  my $app = MyApp::Website->new( url => "http://google.com", name => "Google" );

=head1 SUMMARY

In this recipe, we begin to delve into the wonder of meta-programming.
Some readers may scoff and claim that this is the arena of only the
most twisted Moose developers. Absolutely not! Any sufficiently
twisted developer can benefit greatly from going more meta.

Our goal is to allow each attribute to have a human-readable "label"
attached to it. Such labels would be used when showing data to an end
user. In this recipe we label the C<url> attribute with "The site's
URL" and create a simple method showing how to use that label.

=head1 META-ATTRIBUTE OBJECTS

All the attributes of a Moose-based object are actually objects themselves.
These objects have methods and attributes. Let's look at a concrete example.

  has 'x' => ( isa => 'Int', is => 'ro' );
  has 'y' => ( isa => 'Int', is => 'rw' );

Internally, the metaclass for C<Point> has two L<Moose::Meta::Attribute>
objects. There are several methods for getting meta-attributes out of a
metaclass, one of which is C<get_attribute_list>. This method is called on the
metaclass object.

The C<get_attribute_list> method returns a list of attribute names. You can
then use C<get_attribute> to get the L<Moose::Meta::Attribute> object itself.

Once you have this meta-attribute object, you can call methods on it like
this:

  print $point->meta->get_attribute('x')->type_constraint;
     => Int

To add a label to our attributes there are two steps. First, we need a new
attribute metaclass trait that can store a label for an attribute. Second, we
need to apply that trait to our attributes.

=head1 TRAITS

Roles that apply to metaclasses have a special name: traits. Don't let
the change in nomenclature fool you, B<traits are just roles>.

L<Moose/has> allows you to pass a C<traits> parameter for an
attribute. This parameter takes a list of trait names which are
composed into an anonymous metaclass, and that anonymous metaclass is
used for the attribute.

Yes, we still have lots of metaclasses in the background, but they're
managed by Moose for you.

Traits can do anything roles can do. They can add or refine
attributes, wrap methods, provide more methods, define an interface,
etc. The only difference is that you're now changing the attribute
metaclass instead of a user-level class.

=head1 DISSECTION

We start by creating a package for our trait.

  package MyApp::Meta::Attribute::Trait::Labeled;
  use Moose::Role;

  has label => (
      is        => 'rw',
      isa       => 'Str',
      predicate => 'has_label',
  );

You can see that a trait is just a L<Moose::Role>. In this case, our role
contains a single attribute, C<label>. Any attribute which does this trait
will now have a label.

Next we register our trait with Moose:

  package Moose::Meta::Attribute::Custom::Trait::Labeled;
  sub register_implementation { 'MyApp::Meta::Attribute::Trait::Labeled' }

Moose looks for the C<register_implementation> method in
C<Moose::Meta::Attribute::Custom::Trait::$TRAIT_NAME> to find the full
name of the trait.

Finally, we pass our trait when defining an attribute:

  has url => (
      traits => [qw/Labeled/],
      is     => 'rw',
      isa    => 'Str',
      label  => "The site's URL",
  );

The C<traits> parameter contains a list of trait names. Moose will build an
anonymous attribute metaclass from these traits and use it for this
attribute.

The reason that we can pass the name C<Labeled>, instead of
C<MyApp::Meta::Attribute::Trait::Labeled>, is because of the
C<register_implementation> code we touched on previously.

When you pass a metaclass to C<has>, it will take the name you provide and
prefix it with C<Moose::Meta::Attribute::Custom::Trait::>. Then it calls
C<register_implementation> in the package. In this case, that means Moose ends
up calling
C<Moose::Meta::Attribute::Custom::Trait::Labeled::register_implementation>.

If this function exists, it should return the I<real> trait's package
name. This is exactly what our code does, returning
C<MyApp::Meta::Attribute::Trait::Labeled>. This is a little convoluted, and if
you don't like it, you can always use the fully-qualified name.

We can access this meta-attribute and its label like this:

  $website->meta->get_attribute('url')->label()

  MyApp::Website->meta->get_attribute('url')->label()

We also have a regular attribute, C<name>:

  has name => (
      is  => 'rw',
      isa => 'Str',
  );

Finally, we have a C<dump> method, which creates a human-readable
representation of a C<MyApp::Website> object. It will use an attribute's label
if it has one.

  sub dump {
      my $self = shift;

      my $meta = $self->meta;

      my $dump = '';

      for my $attribute ( map { $meta->get_attribute($_) }
          sort $meta->get_attribute_list ) {

          if (   $attribute->does('MyApp::Meta::Attribute::Trait::Labeled')
              && $attribute->has_label ) {
              $dump .= $attribute->label;
          }

This is a bit of defensive code. We cannot depend on every meta-attribute
having a label. Even if we define one for every attribute in our class, a
subclass may neglect to do so. Or a superclass could add an attribute without
a label.

We also check that the attribute has a label using the predicate we
defined. We could instead make the label C<required>. If we have a label, we
use it, otherwise we use the attribute name:

          else {
              $dump .= $attribute->name;
          }

          my $reader = $attribute->get_read_method;
          $dump .= ": " . $self->$reader . "\n";
      }

      return $dump;
  }

The C<get_read_method> is part of the L<Moose::Meta::Attribute> API. It
returns the name of a method that can read the attribute's value, I<when
called on the real object> (don't call this on the meta-attribute).

=head1 CONCLUSION

You might wonder why you'd bother with all this. You could just hardcode "The
Site's URL" in the C<dump> method. But we want to avoid repetition. If you
need the label once, you may need it elsewhere, maybe in the C<as_form> method
you write next.

Associating a label with an attribute just makes sense! The label is a piece
of information I<about> the attribute.

It's also important to realize that this was a trivial example. You can make
much more powerful metaclasses that I<do> things, as opposed to just storing
some more information. For example, you could implement a metaclass that
expires attributes after a certain amount of time:

   has site_cache => (
       traits        => ['TimedExpiry'],
       expires_after => { hours => 1 },
       refresh_with  => sub { get( $_[0]->url ) },
       isa           => 'Str',
       is            => 'ro',
   );

The sky's the limit!

=begin testing

my $app
    = MyApp::Website->new( url => 'http://google.com', name => 'Google' );
is(
    $app->dump, q{name: Google
The site's URL: http://google.com
}, '... got the expected dump value'
);

=end testing

=cut