| Class-MOP documentation | Contained in the Class-MOP distribution. |
Class::MOP::Class - Class Meta Object
# assuming that class Foo
# has been defined, you can
# use this for introspection ...
# add a method to Foo ...
Foo->meta->add_method( 'bar' => sub {...} )
# get a list of all the classes searched
# the method dispatcher in the correct order
Foo->meta->class_precedence_list()
# remove a method from Foo
Foo->meta->remove_method('bar');
# or use this to actually create classes ...
Class::MOP::Class->create(
'Bar' => (
version => '0.01',
superclasses => ['Foo'],
attributes => [
Class::MOP::Attribute->new('$bar'),
Class::MOP::Attribute->new('$baz'),
],
methods => {
calculate_bar => sub {...},
construct_baz => sub {...}
}
)
);
The Class Protocol is the largest and most complex part of the Class::MOP meta-object protocol. It controls the introspection and manipulation of Perl 5 classes, and it can create them as well. The best way to understand what this module can do is to read the documentation for each of its methods.
Class::MOP::Class is a subclass of Class::MOP::Module.
These methods all create new Class::MOP::Class objects. These
objects can represent existing classes or they can be used to create
new classes from scratch.
The metaclass object for a given class is a singleton. If you attempt to create a metaclass for the same class twice, you will just get the existing object.
This method creates a new Class::MOP::Class object with the given
package name. It accepts a number of options:
An optional version number for the newly created package.
An optional authority for the newly created package.
An optional array reference of superclass names.
An optional hash reference of methods for the class. The keys of the hash reference are method names and values are subroutine references.
An optional array reference of Class::MOP::Attribute objects.
This method works just like Class::MOP::Class->create but it
creates an "anonymous" class. In fact, the class does have a name, but
that name is a unique name generated internally by this module.
It accepts the same superclasses, methods, and attributes
parameters that create accepts.
Anonymous classes are destroyed once the metaclass they are attached to goes out of scope, and will be removed from Perl's internal symbol table.
All instances of an anonymous class keep a special reference to the metaclass object, which prevents the metaclass from going out of scope while any instances exist.
This only works if the instance is based on a hash reference, however.
This method will initialize a Class::MOP::Class object for the
named package. Unlike create, this method will not create a new
class.
The purpose of this method is to retrieve a Class::MOP::Class
object for introspecting an existing class.
If an existing Class::MOP::Class object exists for the named
package, it will be returned, and any options provided will be
ignored!
If the object does not yet exist, it will be created.
The valid options that can be passed to this method are
attribute_metaclass, method_metaclass,
wrapped_method_metaclass, and instance_metaclass. These are all
optional, and default to the appropriate class in the Class::MOP
distribution.
These methods are all related to creating and/or cloning object instances.
This method clones an existing object instance. Any parameters you provide are will override existing attribute values in the object.
This is a convenience method for cloning an object instance, then blessing it into the appropriate package.
You could implement a clone method in your class, using this method:
sub clone {
my ($self, %params) = @_;
$self->meta->clone_object($self, %params);
}
This method changes the class of $instance to the metaclass's class.
You can only rebless an instance into a subclass of its current class. If you pass any additional parameters, these will be treated like constructor parameters and used to initialize the object's attributes. Any existing attributes that are already set will be overwritten.
Before reblessing the instance, this method will call
rebless_instance_away on the instance's current metaclass. This method
will be passed the instance, the new metaclass, and any parameters
specified to rebless_instance. By default, rebless_instance_away
does nothing; it is merely a hook.
This method is used to create a new object of the metaclass's
class. Any parameters you provide are used to initialize the
instance's attributes. A special __INSTANCE__ key can be passed to
provide an already generated instance, rather than having Class::MOP
generate it for you. This is mostly useful for using Class::MOP with
foreign classes which generate instances using their own constructors.
Returns the class name of the instance metaclass. See Class::MOP::Instance for more information on the instance metaclass.
Returns an instance of the instance_metaclass to be used in the
construction of a new instance of the class.
These are a few predicate methods for asking information about the class itself.
This returns true if the class was created by calling Class::MOP::Class->create_anon_class.
This returns true if the class is still mutable.
This returns true if the class has been made immutable.
A class is not pristine if it has non-inherited attributes or if it has any generated methods.
This is a read-write accessor which represents the superclass relationships of the metaclass's class.
This is basically sugar around getting and setting @ISA.
This returns a list of all of the class's ancestor classes. The classes are returned in method dispatch order.
This returns a list based on class_precedence_list but with all
duplicates removed.
This returns a list of all subclasses for this class, even indirect subclasses.
This returns a list of immediate subclasses for this class, which does not include indirect subclasses.
See Method introspection and creation in Class::MOP::Package for methods that operate only on the current class. Class::MOP::Class adds introspection capabilities that take inheritance into account.
This will traverse the inheritance hierarchy and return a list of all the Class::MOP::Method objects for this class and its parents.
This will return a Class::MOP::Method for the specified
$method_name. If the class does not have the specified method, it
returns undef
Unlike get_method, this method will look for the named method in
superclasses.
This will return a list of method names for all of this class's methods, including inherited methods.
This method looks for the named method in the class and all of its parents. It returns every matching method it finds in the inheritance tree, so it returns a list of methods.
Each method is returned as a hash reference with three keys. The keys
are name, class, and code. The code key has a
Class::MOP::Method object as its value.
The list of methods is distinct.
This method returns the first method in any superclass matching the
given name. It is effectively the method that SUPER::$method_name
would dispatch to.
Because Perl 5 does not have a core concept of attributes in classes, we can only return information about attributes which have been added via this class's methods. We cannot discover information about attributes which are defined in terms of "regular" Perl 5 methods.
This will return a Class::MOP::Attribute for the specified
$attribute_name. If the class does not have the specified
attribute, it returns undef.
NOTE that get_attribute does not search superclasses, for that you
need to use find_attribute_by_name.
Returns a boolean indicating whether or not the class defines the named attribute. It does not include attributes inherited from parent classes.
This will return a list of attributes names for all attributes defined in this class.
This will traverse the inheritance hierarchy and return a list of all the Class::MOP::Attribute objects for this class and its parents.
This will return a Class::MOP::Attribute for the specified
$attribute_name. If the class does not have the specified
attribute, it returns undef.
Unlike get_attribute, this attribute will look for the named
attribute in superclasses.
This method accepts either an existing Class::MOP::Attribute
object or parameters suitable for passing to that class's new
method.
The attribute provided will be added to the class.
Any accessor methods defined by the attribute will be added to the class when the attribute is added.
If an attribute of the same name already exists, the old attribute will be removed first.
This will remove the named attribute from the class, and Class::MOP::Attribute object.
Removing an attribute also removes any accessor methods defined by the attribute.
However, note that removing an attribute will only affect future object instances created for this class, not existing instances.
Returns the class name of the attribute metaclass for this class. By default, this is Class::MOP::Attribute.
Making a class immutable "freezes" the class definition. You can no longer call methods which alter the class, such as adding or removing methods or attributes.
Making a class immutable lets us optimize the class by inlining some methods, and also allows us to optimize some methods on the metaclass object itself.
After immutabilization, the metaclass object will cache most informational
methods that returns information about methods or attributes. Methods which
would alter the class, such as add_attribute and add_method, will
throw an error on an immutable metaclass object.
The immutabilization system in Moose takes much greater advantage of the inlining features than Class::MOP itself does.
This method will create an immutable transformer and use it to make the class and its metaclass object immutable.
This method accepts the following options:
These are all booleans indicating whether the specified method(s) should be inlined.
By default, accessors and the constructor are inlined, but not the destructor.
The name of a class which will be used as a parent class for the metaclass object being made immutable. This "trait" implements the post-immutability functionality of the metaclass (but not the transformation itself).
This defaults to Class::MOP::Class::Immutable::Trait.
This is the constructor method name. This defaults to "new".
The name of the method metaclass for constructors. It will be used to generate the inlined constructor. This defaults to "Class::MOP::Method::Constructor".
This is a boolean indicating whether an existing constructor should be replaced when inlining a constructor. This defaults to false.
The name of the method metaclass for destructors. It will be used to generate the inlined destructor. This defaults to "Class::MOP::Method::Denstructor".
This is a boolean indicating whether an existing destructor should be replaced when inlining a destructor. This defaults to false.
Returns a hash of the options used when making the class immutable, including
both defaults and anything supplied by the user in the call to $metaclass->make_immutable. This is useful if you need to temporarily make
a class mutable and then restore immutability as it was before.
Calling this method reverse the immutabilization transformation.
Method modifiers are hooks which allow a method to be wrapped with before, after and around method modifiers. Every time a method is called, its modifiers are also called.
A class can modify its own methods, as well as methods defined in parent classes.
Method modifiers work by wrapping the original method and then replacing it in the class's symbol table. The wrappers will handle calling all the modifiers in the appropriate order and preserving the calling context for the original method.
The return values of before and after modifiers are
ignored. This is because their purpose is not to filter the input
and output of the primary method (this is done with an around
modifier).
This may seem like an odd restriction to some, but doing this allows for simple code to be added at the beginning or end of a method call without altering the function of the wrapped method or placing any extra responsibility on the code of the modifier.
Of course if you have more complex needs, you can use the around
modifier which allows you to change both the parameters passed to the
wrapped method, as well as its return value.
Before and around modifiers are called in last-defined-first-called order, while after modifiers are called in first-defined-first-called order. So the call tree might looks something like this:
before 2
before 1
around 2
around 1
primary
around 1
around 2
after 1
after 2
Of course there is a performance cost associated with method modifiers, but we have made every effort to make that cost directly proportional to the number of modifier features you use.
The wrapping method does its best to only do as much work as it absolutely needs to. In order to do this we have moved some of the performance costs to set-up time, where they are easier to amortize.
All this said, our benchmarks have indicated the following:
simple wrapper with no modifiers 100% slower simple wrapper with simple before modifier 400% slower simple wrapper with simple after modifier 450% slower simple wrapper with simple around modifier 500-550% slower simple wrapper with all 3 modifiers 1100% slower
These numbers may seem daunting, but you must remember, every feature
comes with some cost. To put things in perspective, just doing a
simple AUTOLOAD which does nothing but extract the name of the
method called and return it costs about 400% over a normal method
call.
This wraps the specified method with the supplied subroutine reference. The modifier will be called as a method itself, and will receive the same arguments as are passed to the method.
When the modifier exits, the wrapped method will be called.
The return value of the modifier will be ignored.
This wraps the specified method with the supplied subroutine reference. The modifier will be called as a method itself, and will receive the same arguments as are passed to the method.
When the wrapped methods exits, the modifier will be called.
The return value of the modifier will be ignored.
This wraps the specified method with the supplied subroutine reference.
The first argument passed to the modifier will be a subroutine reference to the wrapped method. The second argument is the object, and after that come any arguments passed when the method is called.
The around modifier can choose to call the original method, as well as what arguments to pass if it does so.
The return value of the modifier is what will be seen by the caller.
This will return a Class::MOP::Class instance for this class.
It should also be noted that Class::MOP will actually bootstrap this module by installing a number of attribute meta-objects into its metaclass.
Stevan Little <stevan@iinteractive.com>
Copyright 2006-2009 by Infinity Interactive, Inc.
This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.
| Class-MOP documentation | Contained in the Class-MOP distribution. |
package Class::MOP::Class; use strict; use warnings; use Class::MOP::Instance; use Class::MOP::Method::Wrapped; use Class::MOP::Method::Accessor; use Class::MOP::Method::Constructor; use Carp 'confess'; use Scalar::Util 'blessed', 'reftype', 'weaken'; use Sub::Name 'subname'; use Devel::GlobalDestruction 'in_global_destruction'; use Try::Tiny; our $VERSION = '0.95'; $VERSION = eval $VERSION; our $AUTHORITY = 'cpan:STEVAN'; use base 'Class::MOP::Module'; # Creation sub initialize { my $class = shift; my $package_name; if ( @_ % 2 ) { $package_name = shift; } else { my %options = @_; $package_name = $options{package}; } ($package_name && !ref($package_name)) || confess "You must pass a package name and it cannot be blessed"; return Class::MOP::get_metaclass_by_name($package_name) || $class->_construct_class_instance(package => $package_name, @_); } # NOTE: (meta-circularity) # this is a special form of _construct_instance # (see below), which is used to construct class # meta-object instances for any Class::MOP::* # class. All other classes will use the more # normal &construct_instance. sub _construct_class_instance { my $class = shift; my $options = @_ == 1 ? $_[0] : {@_}; my $package_name = $options->{package}; (defined $package_name && $package_name) || confess "You must pass a package name"; # NOTE: # return the metaclass if we have it cached, # and it is still defined (it has not been # reaped by DESTROY yet, which can happen # annoyingly enough during global destruction) if (defined(my $meta = Class::MOP::get_metaclass_by_name($package_name))) { return $meta; } # NOTE: # we need to deal with the possibility # of class immutability here, and then # get the name of the class appropriately $class = (ref($class) ? ($class->is_immutable ? $class->_get_mutable_metaclass_name() : ref($class)) : $class); # now create the metaclass my $meta; if ($class eq 'Class::MOP::Class') { $meta = $class->_new($options); } else { # NOTE: # it is safe to use meta here because # class will always be a subclass of # Class::MOP::Class, which defines meta $meta = $class->meta->_construct_instance($options) } # and check the metaclass compatibility $meta->_check_metaclass_compatibility(); Class::MOP::store_metaclass_by_name($package_name, $meta); # NOTE: # we need to weaken any anon classes # so that they can call DESTROY properly Class::MOP::weaken_metaclass($package_name) if $meta->is_anon_class; $meta; } sub _new { my $class = shift; return Class::MOP::Class->initialize($class)->new_object(@_) if $class ne __PACKAGE__; my $options = @_ == 1 ? $_[0] : {@_}; return bless { # inherited from Class::MOP::Package 'package' => $options->{package}, # NOTE: # since the following attributes will # actually be loaded from the symbol # table, and actually bypass the instance # entirely, we can just leave these things # listed here for reference, because they # should not actually have a value associated # with the slot. 'namespace' => \undef, 'methods' => {}, # inherited from Class::MOP::Module 'version' => \undef, 'authority' => \undef, # defined in Class::MOP::Class 'superclasses' => \undef, 'attributes' => {}, 'attribute_metaclass' => ( $options->{'attribute_metaclass'} || 'Class::MOP::Attribute' ), 'method_metaclass' => ( $options->{'method_metaclass'} || 'Class::MOP::Method' ), 'wrapped_method_metaclass' => ( $options->{'wrapped_method_metaclass'} || 'Class::MOP::Method::Wrapped' ), 'instance_metaclass' => ( $options->{'instance_metaclass'} || 'Class::MOP::Instance' ), 'immutable_trait' => ( $options->{'immutable_trait'} || 'Class::MOP::Class::Immutable::Trait' ), 'constructor_name' => ( $options->{constructor_name} || 'new' ), 'constructor_class' => ( $options->{constructor_class} || 'Class::MOP::Method::Constructor' ), 'destructor_class' => $options->{destructor_class}, }, $class; } sub reset_package_cache_flag { (shift)->{'_package_cache_flag'} = undef } sub update_package_cache_flag { my $self = shift; # NOTE: # we can manually update the cache number # since we are actually adding the method # to our cache as well. This avoids us # having to regenerate the method_map. # - SL $self->{'_package_cache_flag'} = Class::MOP::check_package_cache_flag($self->name); } sub _check_metaclass_compatibility { my $self = shift; # this is always okay ... return if ref($self) eq 'Class::MOP::Class' && $self->instance_metaclass eq 'Class::MOP::Instance'; my @class_list = $self->linearized_isa; shift @class_list; # shift off $self->name foreach my $superclass_name (@class_list) { my $super_meta = Class::MOP::get_metaclass_by_name($superclass_name) || next; # NOTE: # we need to deal with the possibility # of class immutability here, and then # get the name of the class appropriately my $super_meta_type = $super_meta->is_immutable ? $super_meta->_get_mutable_metaclass_name() : ref($super_meta); ($self->isa($super_meta_type)) || confess "The metaclass of " . $self->name . " (" . (ref($self)) . ")" . " is not compatible with the " . "metaclass of its superclass, ".$superclass_name . " (" . ($super_meta_type) . ")"; # NOTE: # we also need to check that instance metaclasses # are compatibile in the same the class. ($self->instance_metaclass->isa($super_meta->instance_metaclass)) || confess "The instance metaclass for " . $self->name . " (" . ($self->instance_metaclass) . ")" . " is not compatible with the " . "instance metaclass of its superclass, " . $superclass_name . " (" . ($super_meta->instance_metaclass) . ")"; } } ## ANON classes { # NOTE: # this should be sufficient, if you have a # use case where it is not, write a test and # I will change it. my $ANON_CLASS_SERIAL = 0; # NOTE: # we need a sufficiently annoying prefix # this should suffice for now, this is # used in a couple of places below, so # need to put it up here for now. my $ANON_CLASS_PREFIX = 'Class::MOP::Class::__ANON__::SERIAL::'; sub is_anon_class { my $self = shift; no warnings 'uninitialized'; $self->name =~ /^$ANON_CLASS_PREFIX/o; } sub create_anon_class { my ($class, %options) = @_; my $package_name = $ANON_CLASS_PREFIX . ++$ANON_CLASS_SERIAL; return $class->create($package_name, %options); } # NOTE: # this will only get called for # anon-classes, all other calls # are assumed to occur during # global destruction and so don't # really need to be handled explicitly sub DESTROY { my $self = shift; return if in_global_destruction(); # it'll happen soon anyway and this just makes things more complicated no warnings 'uninitialized'; my $name = $self->name; return unless $name =~ /^$ANON_CLASS_PREFIX/o; # Moose does a weird thing where it replaces the metaclass for # class when fixing metaclass incompatibility. In that case, # we don't want to clean out the namespace now. We can detect # that because Moose will explicitly update the singleton # cache in Class::MOP. my $current_meta = Class::MOP::get_metaclass_by_name($name); return if $current_meta ne $self; my ($serial_id) = ($name =~ /^$ANON_CLASS_PREFIX(\d+)/o); no strict 'refs'; @{$name . '::ISA'} = (); %{$name . '::'} = (); delete ${$ANON_CLASS_PREFIX}{$serial_id . '::'}; Class::MOP::remove_metaclass_by_name($name); } } # creating classes with MOP ... sub create { my ( $class, @args ) = @_; unshift @args, 'package' if @args % 2 == 1; my (%options) = @args; my $package_name = $options{package}; (ref $options{superclasses} eq 'ARRAY') || confess "You must pass an ARRAY ref of superclasses" if exists $options{superclasses}; (ref $options{attributes} eq 'ARRAY') || confess "You must pass an ARRAY ref of attributes" if exists $options{attributes}; (ref $options{methods} eq 'HASH') || confess "You must pass a HASH ref of methods" if exists $options{methods}; my (%initialize_options) = @args; delete @initialize_options{qw( package superclasses attributes methods version authority )}; my $meta = $class->initialize( $package_name => %initialize_options ); $meta->_instantiate_module( $options{version}, $options{authority} ); # FIXME totally lame $meta->add_method('meta' => sub { $class->initialize(ref($_[0]) || $_[0]); }); $meta->superclasses(@{$options{superclasses}}) if exists $options{superclasses}; # NOTE: # process attributes first, so that they can # install accessors, but locally defined methods # can then overwrite them. It is maybe a little odd, but # I think this should be the order of things. if (exists $options{attributes}) { foreach my $attr (@{$options{attributes}}) { $meta->add_attribute($attr); } } if (exists $options{methods}) { foreach my $method_name (keys %{$options{methods}}) { $meta->add_method($method_name, $options{methods}->{$method_name}); } } return $meta; } ## Attribute readers # NOTE: # all these attribute readers will be bootstrapped # away in the Class::MOP bootstrap section sub _attribute_map { $_[0]->{'attributes'} } sub attribute_metaclass { $_[0]->{'attribute_metaclass'} } sub instance_metaclass { $_[0]->{'instance_metaclass'} } sub immutable_trait { $_[0]->{'immutable_trait'} } sub constructor_class { $_[0]->{'constructor_class'} } sub constructor_name { $_[0]->{'constructor_name'} } sub destructor_class { $_[0]->{'destructor_class'} } # Instance Construction & Cloning sub new_object { my $class = shift; # NOTE: # we need to protect the integrity of the # Class::MOP::Class singletons here, so we # delegate this to &construct_class_instance # which will deal with the singletons return $class->_construct_class_instance(@_) if $class->name->isa('Class::MOP::Class'); return $class->_construct_instance(@_); } sub _construct_instance { my $class = shift; my $params = @_ == 1 ? $_[0] : {@_}; my $meta_instance = $class->get_meta_instance(); # FIXME: # the code below is almost certainly incorrect # but this is foreign inheritance, so we might # have to kludge it in the end. my $instance = $params->{__INSTANCE__} || $meta_instance->create_instance(); foreach my $attr ($class->get_all_attributes()) { $attr->initialize_instance_slot($meta_instance, $instance, $params); } # NOTE: # this will only work for a HASH instance type if ($class->is_anon_class) { (reftype($instance) eq 'HASH') || confess "Currently only HASH based instances are supported with instance of anon-classes"; # NOTE: # At some point we should make this official # as a reserved slot name, but right now I am # going to keep it here. # my $RESERVED_MOP_SLOT = '__MOP__'; $instance->{'__MOP__'} = $class; } return $instance; } sub get_meta_instance { my $self = shift; $self->{'_meta_instance'} ||= $self->_create_meta_instance(); } sub _create_meta_instance { my $self = shift; my $instance = $self->instance_metaclass->new( associated_metaclass => $self, attributes => [ $self->get_all_attributes() ], ); $self->add_meta_instance_dependencies() if $instance->is_dependent_on_superclasses(); return $instance; } sub clone_object { my $class = shift; my $instance = shift; (blessed($instance) && $instance->isa($class->name)) || confess "You must pass an instance of the metaclass (" . (ref $class ? $class->name : $class) . "), not ($instance)"; # NOTE: # we need to protect the integrity of the # Class::MOP::Class singletons here, they # should not be cloned. return $instance if $instance->isa('Class::MOP::Class'); $class->_clone_instance($instance, @_); } sub _clone_instance { my ($class, $instance, %params) = @_; (blessed($instance)) || confess "You can only clone instances, ($instance) is not a blessed instance"; my $meta_instance = $class->get_meta_instance(); my $clone = $meta_instance->clone_instance($instance); foreach my $attr ($class->get_all_attributes()) { if ( defined( my $init_arg = $attr->init_arg ) ) { if (exists $params{$init_arg}) { $attr->set_value($clone, $params{$init_arg}); } } } return $clone; } sub rebless_instance { my ($self, $instance, %params) = @_; my $old_metaclass = Class::MOP::class_of($instance); my $old_class = $old_metaclass ? $old_metaclass->name : blessed($instance); $self->name->isa($old_class) || confess "You may rebless only into a subclass of ($old_class), of which (". $self->name .") isn't."; $old_metaclass->rebless_instance_away($instance, $self, %params) if $old_metaclass; my $meta_instance = $self->get_meta_instance(); # rebless! # we use $_[1] here because of t/306_rebless_overload.t regressions on 5.8.8 $meta_instance->rebless_instance_structure($_[1], $self); foreach my $attr ( $self->get_all_attributes ) { if ( $attr->has_value($instance) ) { if ( defined( my $init_arg = $attr->init_arg ) ) { $params{$init_arg} = $attr->get_value($instance) unless exists $params{$init_arg}; } else { $attr->set_value($instance, $attr->get_value($instance)); } } } foreach my $attr ($self->get_all_attributes) { $attr->initialize_instance_slot($meta_instance, $instance, \%params); } $instance; } sub rebless_instance_away { # this intentionally does nothing, it is just a hook } # Inheritance sub superclasses { my $self = shift; my $var_spec = { sigil => '@', type => 'ARRAY', name => 'ISA' }; if (@_) { my @supers = @_; @{$self->get_package_symbol($var_spec)} = @supers; # NOTE: # on 5.8 and below, we need to call # a method to get Perl to detect # a cycle in the class hierarchy my $class = $self->name; $class->isa($class); # NOTE: # we need to check the metaclass # compatibility here so that we can # be sure that the superclass is # not potentially creating an issues # we don't know about $self->_check_metaclass_compatibility(); $self->_superclasses_updated(); } @{$self->get_package_symbol($var_spec)}; } sub _superclasses_updated { my $self = shift; $self->update_meta_instance_dependencies(); } sub subclasses { my $self = shift; my $super_class = $self->name; return @{ $super_class->mro::get_isarev() }; } sub direct_subclasses { my $self = shift; my $super_class = $self->name; return grep { grep { $_ eq $super_class } Class::MOP::Class->initialize($_)->superclasses } $self->subclasses; } sub linearized_isa { return @{ mro::get_linear_isa( (shift)->name ) }; } sub class_precedence_list { my $self = shift; my $name = $self->name; unless (Class::MOP::IS_RUNNING_ON_5_10()) { # NOTE: # We need to check for circular inheritance here # if we are are not on 5.10, cause 5.8 detects it # late. This will do nothing if all is well, and # blow up otherwise. Yes, it's an ugly hack, better # suggestions are welcome. # - SL ($name || return)->isa('This is a test for circular inheritance') } # if our mro is c3, we can # just grab the linear_isa if (mro::get_mro($name) eq 'c3') { return @{ mro::get_linear_isa($name) } } else { # NOTE: # we can't grab the linear_isa for dfs # since it has all the duplicates # already removed. return ( $name, map { $self->initialize($_)->class_precedence_list() } $self->superclasses() ); } } ## Methods { my $fetch_and_prepare_method = sub { my ($self, $method_name) = @_; my $wrapped_metaclass = $self->wrapped_method_metaclass; # fetch it locally my $method = $self->get_method($method_name); # if we dont have local ... unless ($method) { # try to find the next method $method = $self->find_next_method_by_name($method_name); # die if it does not exist (defined $method) || confess "The method '$method_name' was not found in the inheritance hierarchy for " . $self->name; # and now make sure to wrap it # even if it is already wrapped # because we need a new sub ref $method = $wrapped_metaclass->wrap($method, package_name => $self->name, name => $method_name, ); } else { # now make sure we wrap it properly $method = $wrapped_metaclass->wrap($method, package_name => $self->name, name => $method_name, ) unless $method->isa($wrapped_metaclass); } $self->add_method($method_name => $method); return $method; }; sub add_before_method_modifier { my ($self, $method_name, $method_modifier) = @_; (defined $method_name && length $method_name) || confess "You must pass in a method name"; my $method = $fetch_and_prepare_method->($self, $method_name); $method->add_before_modifier( subname(':before' => $method_modifier) ); } sub add_after_method_modifier { my ($self, $method_name, $method_modifier) = @_; (defined $method_name && length $method_name) || confess "You must pass in a method name"; my $method = $fetch_and_prepare_method->($self, $method_name); $method->add_after_modifier( subname(':after' => $method_modifier) ); } sub add_around_method_modifier { my ($self, $method_name, $method_modifier) = @_; (defined $method_name && length $method_name) || confess "You must pass in a method name"; my $method = $fetch_and_prepare_method->($self, $method_name); $method->add_around_modifier( subname(':around' => $method_modifier) ); } # NOTE: # the methods above used to be named like this: # ${pkg}::${method}:(before|after|around) # but this proved problematic when using one modifier # to wrap multiple methods (something which is likely # to happen pretty regularly IMO). So instead of naming # it like this, I have chosen to just name them purely # with their modifier names, like so: # :(before|after|around) # The fact is that in a stack trace, it will be fairly # evident from the context what method they are attached # to, and so don't need the fully qualified name. } sub find_method_by_name { my ($self, $method_name) = @_; (defined $method_name && length $method_name) || confess "You must define a method name to find"; foreach my $class ($self->linearized_isa) { my $method = $self->initialize($class)->get_method($method_name); return $method if defined $method; } return; } sub get_all_methods { my $self = shift; my %methods; for my $class ( reverse $self->linearized_isa ) { my $meta = $self->initialize($class); $methods{$_} = $meta->get_method($_) for $meta->get_method_list; } return values %methods; } sub get_all_method_names { my $self = shift; my %uniq; return grep { !$uniq{$_}++ } map { $self->initialize($_)->get_method_list } $self->linearized_isa; } sub find_all_methods_by_name { my ($self, $method_name) = @_; (defined $method_name && length $method_name) || confess "You must define a method name to find"; my @methods; foreach my $class ($self->linearized_isa) { # fetch the meta-class ... my $meta = $self->initialize($class); push @methods => { name => $method_name, class => $class, code => $meta->get_method($method_name) } if $meta->has_method($method_name); } return @methods; } sub find_next_method_by_name { my ($self, $method_name) = @_; (defined $method_name && length $method_name) || confess "You must define a method name to find"; my @cpl = $self->linearized_isa; shift @cpl; # discard ourselves foreach my $class (@cpl) { my $method = $self->initialize($class)->get_method($method_name); return $method if defined $method; } return; } ## Attributes sub add_attribute { my $self = shift; # either we have an attribute object already # or we need to create one from the args provided my $attribute = blessed($_[0]) ? $_[0] : $self->attribute_metaclass->new(@_); # make sure it is derived from the correct type though ($attribute->isa('Class::MOP::Attribute')) || confess "Your attribute must be an instance of Class::MOP::Attribute (or a subclass)"; # first we attach our new attribute # because it might need certain information # about the class which it is attached to $attribute->attach_to_class($self); my $attr_name = $attribute->name; # then we remove attributes of a conflicting # name here so that we can properly detach # the old attr object, and remove any # accessors it would have generated if ( $self->has_attribute($attr_name) ) { $self->remove_attribute($attr_name); } else { $self->invalidate_meta_instances(); } # get our count of previously inserted attributes and # increment by one so this attribute knows its order my $order = (scalar keys %{$self->_attribute_map}); $attribute->_set_insertion_order($order); # then onto installing the new accessors $self->_attribute_map->{$attr_name} = $attribute; # invalidate package flag here try { local $SIG{__DIE__}; $attribute->install_accessors(); } catch { $self->remove_attribute($attr_name); die $_; }; return $attribute; } sub update_meta_instance_dependencies { my $self = shift; if ( $self->{meta_instance_dependencies} ) { return $self->add_meta_instance_dependencies; } } sub add_meta_instance_dependencies { my $self = shift; $self->remove_meta_instance_dependencies; my @attrs = $self->get_all_attributes(); my %seen; my @classes = grep { not $seen{$_->name}++ } map { $_->associated_class } @attrs; foreach my $class ( @classes ) { $class->add_dependent_meta_instance($self); } $self->{meta_instance_dependencies} = \@classes; } sub remove_meta_instance_dependencies { my $self = shift; if ( my $classes = delete $self->{meta_instance_dependencies} ) { foreach my $class ( @$classes ) { $class->remove_dependent_meta_instance($self); } return $classes; } return; } sub add_dependent_meta_instance { my ( $self, $metaclass ) = @_; push @{ $self->{dependent_meta_instances} }, $metaclass; } sub remove_dependent_meta_instance { my ( $self, $metaclass ) = @_; my $name = $metaclass->name; @$_ = grep { $_->name ne $name } @$_ for $self->{dependent_meta_instances}; } sub invalidate_meta_instances { my $self = shift; $_->invalidate_meta_instance() for $self, @{ $self->{dependent_meta_instances} }; } sub invalidate_meta_instance { my $self = shift; undef $self->{_meta_instance}; } sub has_attribute { my ($self, $attribute_name) = @_; (defined $attribute_name) || confess "You must define an attribute name"; exists $self->_attribute_map->{$attribute_name}; } sub get_attribute { my ($self, $attribute_name) = @_; (defined $attribute_name) || confess "You must define an attribute name"; return $self->_attribute_map->{$attribute_name} # NOTE: # this will return undef anyway, so no need ... # if $self->has_attribute($attribute_name); #return; } sub remove_attribute { my ($self, $attribute_name) = @_; (defined $attribute_name) || confess "You must define an attribute name"; my $removed_attribute = $self->_attribute_map->{$attribute_name}; return unless defined $removed_attribute; delete $self->_attribute_map->{$attribute_name}; $self->invalidate_meta_instances(); $removed_attribute->remove_accessors(); $removed_attribute->detach_from_class(); return $removed_attribute; } sub get_attribute_list { my $self = shift; keys %{$self->_attribute_map}; } sub get_all_attributes { my $self = shift; my %attrs = map { %{ $self->initialize($_)->_attribute_map } } reverse $self->linearized_isa; return values %attrs; } sub find_attribute_by_name { my ($self, $attr_name) = @_; foreach my $class ($self->linearized_isa) { # fetch the meta-class ... my $meta = $self->initialize($class); return $meta->get_attribute($attr_name) if $meta->has_attribute($attr_name); } return; } # check if we can reinitialize sub is_pristine { my $self = shift; # if any local attr is defined return if $self->get_attribute_list; # or any non-declared methods for my $method ( map { $self->get_method($_) } $self->get_method_list ) { return if $method->isa("Class::MOP::Method::Generated"); # FIXME do we need to enforce this too? return unless $method->isa( $self->method_metaclass ); } return 1; } ## Class closing sub is_mutable { 1 } sub is_immutable { 0 } sub immutable_options { %{ $_[0]{__immutable}{options} || {} } } sub _immutable_options { my ( $self, @args ) = @_; return ( inline_accessors => 1, inline_constructor => 1, inline_destructor => 0, debug => 0, immutable_trait => $self->immutable_trait, constructor_name => $self->constructor_name, constructor_class => $self->constructor_class, destructor_class => $self->destructor_class, @args, ); } sub make_immutable { my ( $self, @args ) = @_; if ( $self->is_mutable ) { $self->_initialize_immutable( $self->_immutable_options(@args) ); $self->_rebless_as_immutable(@args); return $self; } else { return; } } sub make_mutable { my $self = shift; if ( $self->is_immutable ) { my @args = $self->immutable_options; $self->_rebless_as_mutable(); $self->_remove_inlined_code(@args); delete $self->{__immutable}; return $self; } else { return; } } sub _rebless_as_immutable { my ( $self, @args ) = @_; $self->{__immutable}{original_class} = ref $self; bless $self => $self->_immutable_metaclass(@args); } sub _immutable_metaclass { my ( $self, %args ) = @_; if ( my $class = $args{immutable_metaclass} ) { return $class; } my $trait = $args{immutable_trait} = $self->immutable_trait || confess "no immutable trait specified for $self"; my $meta = $self->meta; my $meta_attr = $meta->find_attribute_by_name("immutable_trait"); my $class_name; if ( $meta_attr and $trait eq $meta_attr->default ) { # if the trait is the same as the default we try and pick a # predictable name for the immutable metaclass $class_name = 'Class::MOP::Class::Immutable::' . ref($self); } else { $class_name = join '::', 'Class::MOP::Class::Immutable::CustomTrait', $trait, 'ForMetaClass', ref($self); } return $class_name if Class::MOP::is_class_loaded($class_name); # If the metaclass is a subclass of CMOP::Class which has had # metaclass roles applied (via Moose), then we want to make sure # that we preserve that anonymous class (see Fey::ORM for an # example of where this matters). my $meta_name = $meta->is_immutable ? $meta->_get_mutable_metaclass_name : ref $meta; my $immutable_meta = $meta_name->create( $class_name, superclasses => [ ref $self ], ); Class::MOP::load_class($trait); for my $meth ( Class::MOP::Class->initialize($trait)->get_all_methods ) { my $meth_name = $meth->name; if ( $immutable_meta->find_method_by_name( $meth_name ) ) { $immutable_meta->add_around_method_modifier( $meth_name, $meth->body ); } else { $immutable_meta->add_method( $meth_name, $meth->clone ); } } $immutable_meta->make_immutable( inline_constructor => 0, inline_accessors => 0, ); return $class_name; } sub _remove_inlined_code { my $self = shift; $self->remove_method( $_->name ) for $self->_inlined_methods; delete $self->{__immutable}{inlined_methods}; } sub _inlined_methods { @{ $_[0]{__immutable}{inlined_methods} || [] } } sub _add_inlined_method { my ( $self, $method ) = @_; push @{ $self->{__immutable}{inlined_methods} ||= [] }, $method; } sub _initialize_immutable { my ( $self, %args ) = @_; $self->{__immutable}{options} = \%args; $self->_install_inlined_code(%args); } sub _install_inlined_code { my ( $self, %args ) = @_; # FIXME $self->_inline_accessors(%args) if $args{inline_accessors}; $self->_inline_constructor(%args) if $args{inline_constructor}; $self->_inline_destructor(%args) if $args{inline_destructor}; } sub _rebless_as_mutable { my $self = shift; bless $self, $self->_get_mutable_metaclass_name; return $self; } sub _inline_accessors { my $self = shift; foreach my $attr_name ( $self->get_attribute_list ) { $self->get_attribute($attr_name)->install_accessors(1); } } sub _inline_constructor { my ( $self, %args ) = @_; my $name = $args{constructor_name}; if ( $self->has_method($name) && !$args{replace_constructor} ) { my $class = $self->name; warn "Not inlining a constructor for $class since it defines" . " its own constructor.\n" . "If you are certain you don't need to inline your" . " constructor, specify inline_constructor => 0 in your" . " call to $class->meta->make_immutable\n"; return; } my $constructor_class = $args{constructor_class}; Class::MOP::load_class($constructor_class); my $constructor = $constructor_class->new( options => \%args, metaclass => $self, is_inline => 1, package_name => $self->name, name => $name, ); if ( $args{replace_constructor} or $constructor->can_be_inlined ) { $self->add_method( $name => $constructor ); $self->_add_inlined_method($constructor); } } sub _inline_destructor { my ( $self, %args ) = @_; ( exists $args{destructor_class} && defined $args{destructor_class} ) || confess "The 'inline_destructor' option is present, but " . "no destructor class was specified"; if ( $self->has_method('DESTROY') && ! $args{replace_destructor} ) { my $class = $self->name; warn "Not inlining a destructor for $class since it defines" . " its own destructor.\n"; return; } my $destructor_class = $args{destructor_class}; Class::MOP::load_class($destructor_class); return unless $destructor_class->is_needed($self); my $destructor = $destructor_class->new( options => \%args, metaclass => $self, package_name => $self->name, name => 'DESTROY' ); if ( $args{replace_destructor} or $destructor->can_be_inlined ) { $self->add_method( 'DESTROY' => $destructor ); $self->_add_inlined_method($destructor); } } 1; __END__