NAME
Cache::Memcached::AnyEvent - AnyEvent Compatible Memcached Client
SYNOPSIS
use Cache::Memcached::AnyEvent;
my $memd = Cache::Memcached::AnyEvent->new({
servers => [ '127.0.0.1:11211' ],
compress_threshold => 10_000,
namespace => 'myapp.',
});
$memd->get( $key, sub {
my ($value) = @_;
warn "got $value for $key";
});
$memd->disconnect();
# use ketama algorithm instead of the traditional one
my $memd = Cache::Memcached::AnyEvent->new({
...
selector_class => 'Ketama',
# or, selector => $object
});
# use binary protocol instead of text
my $memd = Cache::Memcached::AnyEvent->new({
...
protocol_class => 'Binary',
# or, protocol => $object,
});
DESRIPTION
WARNING: BETA QUALITY CODE!
This module implements the memcached protocol as a AnyEvent consumer, and it implments both for text and binary protocols.
RATIONALE
There's another alternative AnyEvent memcached client, AnyEvent::Memcached which is perfectly fine, and I have nothing against you using that module. I just have some specific itches to scratch:
- Prerequisites
-
This module, Cache::Memcached::AnyEvent, requires the bare minimum prerequisites to install. AnyEvent::Memcached requires AnyEvent::Connection and Object::Event ;) Those modules are fine, I just don't use them, so I don't want them.
- Binary Protocol
-
I was in the mood to implement the binary protocol. I don't believe it's a requirement to do anything, so this is purely a whim. There's nothing that requires binary protocol in the wild, so it has no practactical advantages. I just wanted to implement it :)
- Cache::Memcached Interface
-
In general, this module follows the interface of Cache::Memcached.
So choose according to your needs. If you for some reason don't want AnyEvent::Connection and Object::Event, want a binary protocol, and like to stick with Cache::Memcached interface (relatively speaking), then use this module. Otherwise, read the docs for each module, and choose the one that fits your needs.
METHODS
All methods interacting with a memcached server which can take a callback function can also take a condvar instead. For example,
$memd->get( "foo", sub {
my $value = shift;
} );
is equivalent to
my $cv = AE::cv {
my $value = $_[0]->recv;
};
$memd->get( "foo", $cv );
# optionally, call $cv->recv here.
new(%args)
- auto_reconnect => $max_attempts
-
Set to 0 to disable auto-reconnecting
- compress_threshold => $number
- selector => $object
-
The selector is an object responsible for selecting the appropriate Memcached server to store a particular key. This object MUST implement the following methods:
$object->set_server( @servernames ); my $handle = $object->get_handle( $key );
By default if this argument is not specified, a selector object will automatically be created using the value of the
selector_class
argument. - selector_class => $class_name_or_fragment
-
Specifies the selector class to be instantiated. The default value is "Traditional".
If the class name is preceded by a single '+', then that class name with the '+' removed will be used as the class name. Otherwise, the prefix "Cache::Memcached::AnyEvent::Selector::" will be added to the value ("Traditional" would be transformed to "Cache::Memcached::AnyEvent::Selector::Traditional")
- namespace => $namespace
- procotol => $object
-
The protocol is an object responsible for handling the actual talking to the memcached servers. This object MUST implement all of the memcached interface supported by Cache::Memcached::AnyEvent
By default if this argument is not specified, a protocol object will automatically be created using the value of the
protocol_class
argument. - protocol_class => $classname
-
Specifies the protocol class to be instantiated. The default value is "Text".
If the class name is preceded by a single '+', then that class name with the '+' removed will be used as the class name. Otherwise, the prefix "Cache::Memcached::AnyEvent::Protocol::" will be added to the value ("Text" would be transformed to "Cache::Memcached::AnyEvent::Protocol::Text")
- reconnect_delay => $seconds
-
Amount of time to wait between reconnect attempts
- servers => \@servers
-
List of servers to use.
%args
can also be a hashref.
auto_reconnect([$bool]);
Get/Set auto_reconnect flag.
add($key, $value[, $exptime, $noreply], $cb->($rc))
add_server( $host_port )
append($key, $value, $cb->($rc))
connect()
Explicitly connects to each server given. You DO NOT need to call this explicitly.
decr($key, $delta[, $initial], $cb->($value))
delete($key, $cb->($rc))
disconnect()
flush_all()
get($key, $cb->($value))
get_handle( $host_port )
get_multi(\@keys, $cb->(\%values));
incr($key, $delta[, $initial], $cb->($value))
prepend($key, $value, $cb->($rc));
protocol($object)
replace($key, $value[, $exptime, $noreply], $cb->($rc))
remove($key, $cb->($rc))
Alias to delete
servers()
set($key, $value[, $exptime, $noreply], $cb->($rc))
stats($cmd, $cb->(\%stats))
version( $cb->(\%result) )
TODO
CONTRIBUTING
Contribution is welcome, but please make sure to follow this guideline:
- Please send changes AND tests.
-
In case of changes that supposedly fixes incorrect behavior, you MUST provide me with a failing test case. How should we know if you were hallucinating or just plain stupid if you can't reproduce it yourself?
- Please send code, not an essay.
-
Please don't waste time trying to explain in a 10K email for stuff you can write in 10 lines of code.
I'm sorry, but I'm usually not interested in your opinion until I see some code. Writing essays to describe a software problem is stupid.
- Please avoid emailing patches.
-
If at all possible, please use github pull requests instead of emailing me patches. Patches are stupid.
Please refer to the META.yml file for the repository location.
So to summarize, contribution is welcome, but please don't be stupid. If you can't follow the guideline, you will not be taken seriously.
AUTHOR
Daisuke Maki <daisuke@endeworks.jp>
LICENSE
This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.
See http://www.perl.com/perl/misc/Artistic.html