Linux Certif

Toute la documentation sur la certification Linux LPI

perl5i::Meta.3pm

NAME

perl5i::Meta - The perl5i meta object

SYNOPSIS

     use perl5i;
 
     my $id      = $object->mo->id;
     my $class   = $object->mc->class;
     my $tainted = $object->mo->is_tainted;
     ...and so on...
 
 

DESCRIPTION

Each object has a meta object which can be used to describe and sometimes alter the object. This is for things which are common to *all* objects. For example, "$obj->mc->class" to get the object's class. "@ISA = $obj->mc->ISA" to get an object's parents. And so on.

Why a meta object?

Why not just stick these methods in UNIVERSAL? They'd clash with user-space methods. For example, if an existing class has its own "id()" method it would likely clash with what our "id()" method does. You want to guarantee that every object responds to these meta methods the same way so there's no second-guessing.

Meta Instance vs Meta Class

Each object has a meta object for their instance, accessable with "$obj->mo" and also a meta object for their class, accessable with "$obj->mc". The meta instance can do most everything the meta class can, mc is provided mostly for disambiguation.

The problem is this:

     my $thing = "Foo";
     say $thing->mo->class;
 
 

In perl5i, everything is an object. Do you want the class of $thing or do you want to treat $thing as a class name? Its ambiguous. So to disambiguate, use "$thing->mc" when you mean $thing to be a class name and "$thing->mo" when you mean it to be an object.

For example, when writing a method which could be a class or could be an object be sure to use "$proto->mc->class" to get the class name.

     sub my_method {
         my $proto = shift;  # could be an object, could be a class name
         my $class = $proto->mc->class;
         ....
     }
 
 

METHODS

id

     my $id = $object->mo->id;
 
 

Returns an identifer for $object.

The identifier is guaranteed to be:

   * unique to the object for the life of the process
   * a true value
   * independent of changes to the object's contents
 
 

class

     my $class = $object->mo->class;
     my $class = $class->mc->class;
 
 

Returns the class of the $object or $class.

ISA

     my @ISA = $object->mo->ISA;
     my @ISA = $class->mc->ISA;
 
 

Returns the immediate parents of the $class or $object.

Essentially equivalent to:

     no strict 'refs';
     my @ISA = @{$class.'::ISA'};
 
 

linear_isa

     my @isa = $class->mc->linear_isa();
     my @isa = $object->mo->linear_isa();
 
 

Returns the entire inheritance tree of the $class or $object as a list in the order it will be searched for method inheritance.

This list includes the $class itself and includes UNIVERSAL. For example:

     package Child;
     use parent qw(Parent);
 
     # Child, Parent, UNIVERSAL
     my @isa = Child->mo->linear_isa();
 
 

super

     my @return = $class->mc->super(@args);
     my @return = $object->mo->super(@args);
 
 

Call the parent of $class/$object's implementation of the current method.

Equivalent to "$object->SUPER::method(@args)" but based on the class of the $object rather than the class in which the current method was declared.

is_tainted

     my $is_tainted = $object->mo->is_tainted;
 
 

Returns true if the $object is tainted.

Only scalars can be tainted, so objects generally return false.

String and numerically overloaded objects will check against their overloaded versions.

taint

     $object->mo->taint;
 
 

Taints the $object.

Normally only scalars can be tainted, this will throw an exception on anything else.

Tainted, string overloaded objects will cause this to be a no-op.

An object can override this method if they have a means of tainting themselves. Generally this is applicable to string or numeric overloaded objects who can taint their overloaded value.

untaint

     $object->mo->untaint;
 
 

Untaints the $object.

Normally objects cannot be tainted, so it is a no op on anything but a scalar.

Tainted, string overloaded objects will throw an exception.

An object can override this method if they have a means of untainting themselves. Generally this is applicable to string or numeric overloaded objects who can untaint their overloaded value.

reftype

     my $reftype = $object->mo->reftype;
 
 

Returns the underlying reference type of the $object.

checksum

     my $checksum = $object->mo->checksum;
     my $md5    = $object->mo->checksum( algorithm => 'md5' );
     my $base64 = $object->mo->checksum( format => 'base64' );
 
 

Get a digest of the object's contents, taking its class into account.

Two different objects can have the same checksum if their contents are identical. Likewise, a single object can have different checksums throughout its life cycle if it's mutable. This means its checksum will change if its internal state changes.

For example,

     $obj->mo->checksum( format => 'base64', algorithm => 'md5' );
 
 

options

algorithm

The checksum algorithm. Can be "sha1" and "md5".

Defaults to sha1.

format

The character set of the checksum, can be "hex", "base64", or "binary".

Defaults to hex.

is_equal

     $object->mo->is_equal($other_object)
 
 

Assess whether something is equal to something else, recurring over deep data structures and treating overloaded objects as numbers or strings when appropriate.

Examples:

     my $prices = { chair => 50, table => 300 };
     my $other  = { chair => 50, table => [250, 255] };
 
     say "They are equal" if $prices->mo->is_equal($other);
 
 
     my $uri = URI->new("http://www.perl.org");
     $uri->mo->is_equal("http://www.perl.org") # True
 
 

perl

     my $dump = $object->mo->perl;
 
 

Dumps the contents of the $object as Perl in a string, like Data::Dumper.

dump

     my $dump = $object->mo->dump( format => $format );
 
 

Dumps the contents of the $object as a string in whatever format you like.

Possible formats are yaml, json and perl.

$format defaults to ``perl'' which is equivalent to "$object->mo->perl".