Exploiting the newer perl to improve your plugins

1H.com

1H.com

Exploiting the Newer Perl
to Improve Your Plugins

Marian Marinov - mm@1h.com
Co-founder & CEO of 1H Ltd.

AGENDA

Building feature rich plug-ins the easy way

Queuing for the masses

Cache me please

$ $$ $$$ $$ $

If you rate my survey, I'll hook you up with $20 cPCache $$$.

Go to this address to take the survey:
http://go.cpanel.net/b27

and come up to the podium once you've completed it.

Plug-ins creation

jQuery

JavaScript::Packer

JSON::XS

cPanel plug-ins with jQuery

#!/usr/bin/perl -w
use strict;
use JQuery::Demo;
use JQuery::CSS;
use CGI;

package main;
my $tester = new JQuery::Demo;
$tester->run;

package JQuery::Demo;
use JQuery::Tabs;


sub start {
my $my = shift;
my $q = new CGI;
$my->{info}{TITLE} = "Tabs";
my $jquery = $my->{jquery};
my @tabs = ("tab 1","tab 2","tab 3","tab 4");
my @texts = ("line 1","line 2","line 3","line4");


my $tab = JQuery::Tabs->new(id => 'myTab',
addToJQuery => $jquery,
tabs => \@tabs,
remote => 'true',
remoteProgram => 'jquery_tabs_results.pl',
rm => 'myMode',
spinner => 1 );
my $html = $tab->HTML;
$my->{info}{BODY} = qq[<h1>START OF TAB EXAMPLE</h1>$html</div>END OF EXAMPLE</h1>];
}

JavaScript::Packer

$ wc -l uncompressed.js
1078 uncompressed.js
$ ls -l uncompressed.js
-rw-r--r-- 1 hackman hackman 31609 Oct 8 00:43 uncompressed.js
$ ./minify-js.pl
$ ls -l compressed.js
-rw-rw-r-- 1 hackman hackman 15847 Oct 8 00:43 compressed.js
$ wc -l compressed.js
0 compressed.js

JavaScript::Packer

#!/usr/bin/perl -w
use strict;
use JavaScript::Packer;
my $packer = JavaScript::Packer->init();
open( UNCOMPRESSED, 'uncompressed.js' );
open( COMPRESSED, '>compressed.js' );
my $js = join( '', <UNCOMPRESSED> );
$packer->minify( \$js, { compress => 'best' });
print COMPRESSED $js;
close(UNCOMPRESSED);
close(COMPRESSED);

Why JSON and why JSON::XS

We need to transfer data

XML is EVIL

The best format with the least overhead

The XS version is obviously faster

Example with JSON

#!/usr/bin/perl -w
use strict;
use JSON::XS;
my %res = ( 'a' => 'line1', 'b' => 'line2' );
my $json = JSON::XS->new->ascii->pretty->allow_nonref;
print $json->encode(\%res);

$ ./json-ex.pl
{"a":"line1","b":"line2"}

Why queuing is helpful

Performance

No need to handle your own forks

Queing slow commands

The data is "HOT"

Concurrency

"HOT" connections

No need to think about the position in a file

Security

You have to encode and check the data sent to the functions

Introduce Queuing

Gearman - http://gearman.org

RabbitMQ - http://www.rabbitmq.com

Gearman Architecture

Server - handles the queues

Client - adds requests to the queue

Worker - handles requests from the queue

Worker

Client

gearmand

Web App

Daemon

Gearman client

#!/usr/bin/perl -w
use strict;
use Gearman::Client;
my $client = Gearman::Client->new;
$client->job_servers('127.0.0.1:4730');
$str = "my client has called";
eval {
local $SIG{ALRM} = sub { die 'timeout'; };
alarm(3);
$client->dispatch_background('log_me', $str);
alarm(0);
};

Gearman worker

#!/usr/bin/perl -w
use strict;
use Gearman::Worker;
use POSIX qw(strftime);
$logfile = '/home/hackman/test.log';
open LOG, '>>', $logfile;
select((select(LOG), $| = 1)[0]);
my $logger = sub { ... };
my $w = Gearman::Worker->new;
$w->job_servers('127.0.0.1:4730');
$w->register_function( log_me => $logger );
$w->work while 1;

Gearman setup

# make && make install
# gearmand -d -u gearman -p 4730

# gearman_status.pl
function name queue exec worker
log_me 1 1 1
check_oneh 0 0 1
gather_child 154 72 72

RabbitMQ architecture

RabbitMQ

Queue

Exchange

Server

Server

Queue

Server

RabbitMQ architecture

RabbitMQ

AMQP

XMPP

STOMP

HTTP

AMQP

XMPP

STOMP

HTTP

RabbitMQ setup

# yum install erlang
# rpm -Uvh rabbitmq-server-2.6.1-1.noarch.rpm

# rabbitmq-server -detached
# rabbitmqctl status
Status of node rabbit@gamelon ...
[{pid,6829}, {running_applications,
[{rabbit,"RabbitMQ","2.6.1"},
....

RabbitMQ client

#!/usr/bin/perl -w
use strict;
use Net::RabbitMQ;
my $mq = Net::RabbitMQ->new();
$mq->connect("localhost", { user => "guest", password => "guest" });
$mq->channel_open(1);
$mq->publish(1, "testqueue", "Hi there!");
$mq->disconnect();

RabbitMQ server

#!/usr/bin/perl -w
use strict; use Data::Dumper ; use Net::RabbitMQ;
my $mq = Net::RabbitMQ->new();
$mq->connect("localhost",{user=>"guest",password=>"guest"});
$mq->channel_open(1);
$mq->exchange_declare(1, 'my_x',{auto_delete => 0});
$mq->queue_declare(1, 'testqueue',{auto_delete => 0});
$mq->queue_bind(1, 'testqueue', 'my_x', 'foo');
while(1){
my $hashref = $mq->get(1, 'testqueue');
print Dumper($hashref) if (defined($hashref));
}

Caching

Why cache ?

Cache::memcached

Cache::Redis - written especially for this talk :)

Cache::Memcached

memcached

CGI

CGI

CGI

CGI

CGI

Questions

? ? ? ?
? ? ? ? ? ?
? ? ? ? ? ? ?
? ? ? ? ? ?
? ?

Thank you

Please visit us at Booth 23

Marian Marinov - mm@1h.com
Co-founder & CIO at 1H Ltd.

1H.com

1H.com

Exploiting the newer perl to improve your plugins

More Related Content

What's hot

Similar to Exploiting the newer perl to improve your plugins

More from Marian Marinov

Recently uploaded

Exploiting the newer perl to improve your plugins

Editor's Notes