| Name : perl-Class-Null
| |
| Version : 2.110.730
| Vendor : obs://build_opensuse_org/devel:languages:perl
|
| Release : 1.6
| Date : 2024-03-12 18:02:11
|
| Group : Unspecified
| Source RPM : perl-Class-Null-2.110.730-1.6.src.rpm
|
| Size : 0.03 MB
| |
| Packager : (none)
| |
| Summary : Implements the Null Class design pattern
|
Description :
This class implements the Null Class design pattern.
Suppose that methods in your object want to write log messages to a log
object. The log object is possibly stored in a slot in your object and can
be accessed using an accessor method:
package MyObject;
use base \'Class::Accessor\';
__PACKAGE__->mk_accessors(qw(log));
sub do_it {
my $self = shift;
$self->log->log(level => \'debug\', message => \'starting to do it\');
...
$self->log->log(level => \'debug\', message => \'still doing it\');
...
$self->log->log(level => \'debug\', message => \'finished doing it\');
}
The log object simply needs to have a \'log()\' method that accepts two named
parameters. Any class defining such a method will do, and \'Log::Dispatch\'
fulfills that requirement while providing a lot of flexibility and
reusability in handling the logged messages.
You might want to log messages to a file:
use Log::Dispatch;
my $dispatcher = Log::Dispatch->new;
$dispatcher->add(Log::Dispatch::File->new(
name => \'file1\',
min_level => \'debug\',
filename => \'logfile\'));
my $obj = MyObject->new(log => $dispatcher);
$obj->do_it;
But what happens if we don\'t define a log object? Your object\'s methods
would have to check whether a log object is defined before calling the
\'log()\' method. This leads to lots of unwieldy code like
sub do_it {
my $self = shift;
if (defined (my $log = $self->log)) {
$log->log(level => \'debug\', message => \'starting to do it\');
}
...
if (defined (my $log = $self->log)) {
$log->log(level => \'debug\', message => \'still doing it\');
}
...
if (defined (my $log = $self->log)) {
$log->log(level => \'debug\', message => \'finished doing it\');
}
}
The proliferation of if-statements really distracts from the actual call to
\'log()\' and also distracts from the rest of the method code. There is a
better way. We can ensure that there is always a log object that we can
call \'log()\' on, even if it doesn\'t do very much (or in fact, anything at
all).
This object with null functionality is what is called a null object. We can
create the object the usual way, using the \'new()\' constructor, and call
any method on it, and all methods will do the same - nothing. (Actually, it
always returns the same \'Class::Null\' singleton object, enabling method
chaining.) It\'s effectively a catch-all object. We can use this class with
our own object like this:
package MyObject;
use Class::Null;
sub init {
my $self = shift;
...
$self->log(Class::Null->new);
...
}
sub do_it {
my $self = shift;
$self->log->log(level => \'debug\', message => \'starting to do it\');
...
$self->log->log(level => \'debug\', message => \'still doing it\');
...
$self->log->log(level => \'debug\', message => \'finished doing it\');
}
This is only one example of using a null class, but it can be used whenever
you want to make an optional helper object into a mandatory helper object,
thereby avoiding unnecessarily complicated checks and preserving the
transparency of how your objects are related to each other and how they
call each other.
Although \'Class::Null\' is exceedingly simple it has been made into a
distribution and put on CPAN to avoid further clutter and repetitive
definitions.
|
RPM found in directory: /packages/linux-pbone/ftp5.gwdg.de/pub/opensuse/repositories/devel:/languages:/perl/openSUSE_Tumbleweed/noarch |
