GSP
Quick Navigator

Search Site

Unix VPS
A - Starter
B - Basic
C - Preferred
D - Commercial
MPS - Dedicated
Previous VPSs
* Sign Up! *

Support
Contact Us
Online Help
Handbooks
Domain Status
Man Pages

FAQ
Virtual Servers
Pricing
Billing
Technical

Network
Facilities
Connectivity
Topology Map

Miscellaneous
Server Agreement
Year 2038
Credits
 

USA Flag

 

 

Man Pages
Authen::PluggableCaptcha(3) User Contributed Perl Documentation Authen::PluggableCaptcha(3)

Authen::PluggableCaptcha - A pluggable Captcha framework for Perl

IMPORTANT-- the .03 release is incompatible with earlier versions. Most notably: all external hooks for hash mangling have been replaced with object methods ( ie: $obj->{'__Challenge'} is now $obj->challenge ) and keyword arguments expecting a class name have the word '_class' as a suffix.

Authen::PluggableCaptcha is a framework for creating Captchas , based on the idea of creating Captchas with a plugin architecture.

The power of this module is that it creates Captchas in the sense that a programmer writes Perl modules-- not just in the sense that a programmer calls a Captcha library for display.

The essence of a Captcha has been broken down into three components: KeyManager , Challenge and Render -- all of which programmers now have full control over. Mix and match existing classes or create your own. Authen::PluggableCaptcha helps you make your own captcha tests -- and it helps you do it fast.

The KeyManager component handles creating & validatiing keys that are later used to uniquely identify a CAPTCHA. By default the KeyManager uses a time-based key system, but it can be trivially extended to integrate with a database and make single-use keys.

The Challenge component maps a key to a set of instructions, a user prompt , and a correct response.

The render component is used to display the challenge - be it text, image or sound.

  use Authen::PluggableCaptcha;
  use Authen::PluggableCaptcha::Challenge::TypeString;
  use Authen::PluggableCaptcha::Render::Image::Imager;

  # create a new captcha for your form
  my $captcha= Authen::PluggableCaptcha->new( 
    type=> "new", 
    seed=> $session->user->seed , 
    site_secret=> $MyApp::Config::site_secret 
  );
  my $captcha_publickey= $captcha->get_publickey();
  
  # image captcha?  create an html link to your captcha script with the public key
  my $html= qq|<img src="/path/to/captcha.pl?captcha_publickey=${captcha_publickey}"/>|;
  
  # image captcha?  render it
  my $existing_publickey= 'a33d8ce53691848ee1096061dfdd4639_1149624525';
  my $existing_publickey = $apr->param('captcha_publickey');
  my $captcha= Authen::PluggableCaptcha->new( 
    type=> 'existing' , 
    publickey=> $existing_publickey , 
    seed=> $session->user->seed , 
    site_secret=> $MyApp::Config::site_secret 
  );

  # save it as a file
  my $as_string= $captcha->render( 
    challenge_class=> 'Authen::PluggableCaptcha::Challenge::TypeString', 
    render_class=>'Authen::PluggableCaptcha::Render::Image::Imager' ,  
    format=>'jpeg' 
  );
  open(WRITE, ">test.jpg");
  print WRITE $as_string;
  close(WRITE);

  # or serve it yourself
  $r->add_header('Content Type: image/jpeg');
  $r->print( $as_string );
  
  # wait, what if we want to validate the captcha first?
  my $captcha= Authen::PluggableCaptcha->new( 
    type=> 'existing' , 
    publickey=> $apr->param('captcha_publickey'), 
    seed=> $session->user->seed , 
    site_secret= $MyApp::Config::site_secret 
  );
  if ( !$captcha->validate_response( user_response=> $apr->param('captcha_response') ) ) {
          my $reason= $captcha->get_error('validate_response');
          die "could not validate captcha because: ${reason}.";
  };

in the above example, $captcha->new just configures the captcha. $captcha->render actually renders the image. if the captcha is expired (too old by the default configuration) , the default expired captcha routine from the plugin will take place better yet, handle all the timely and ip/request validation in the application logic. the timeliness just makes someone answer a captcha 1x every 5minutes, but doesn't prevent re/mis use

render accepts a 'render_class' argument that will internally dispatch the routines to a new instance of that class.

using this method, multiple renderings and formats can be created using a single key and challenge.

Authen::PluggableCaptcha is a fully modularized and extensible system for making Pluggable Catpcha (Completely Automated Public Turing Test to Tell Computers and Humans Apart) tests.

Pluggable? All Captcha objects are instantiated and interfaced via the main module, and then manipulated to require various submodules as plug-ins.

Authen::PluggableCaptcha borrows from the functionality in Apache::Session::Flex

KeyManager

  Consolidates functionality previously found in KeyGenerator and KeyValidator

  Generates , parses and validates publickeys which are used to validate and create captchas
  Default is Authen::PluggableCaptcha::KeyManager , which makes a key %md5%_%time% and performs no additional checking

  A subclass is highly recommended.
  Subclasses can contain a regex or a bunch of DB interaction stuff to ensure a key is used only one time per ip address

Challenge

  simply put, a challenge is a test.  
  challenges internally require a ref to a KeyManager instance , it then maps that instance via it's own facilities into a test to render or validate
  a challege generates 3 bits of text: 
        instructions
        user_prompt
        correct_response

  a visual captcha would have user_prompt and correct_response as the same.  
  a text logic puzzle would not.

Render

  the rendering of a captcha for presentation to a user.
  This could be an image, sound, block of (obfuscated?) html or just plain text

Current CPAN captcha modules all exhibit one or more of the following traits:
- the module is tied heavily into a given image rendering library
- the module only supports a single style of an image Catpcha
- the module renders/saves the image to disk

I wanted a module that works in a clustered environment, could be easily extended / implemented with the following design requirements:

1 challenges are presented by a public_key
2 a seed (sessionID ?) + a server key (siteSecret) hash together to create a public key
3 the public_key is handled by its own module which can be subclassed and replaced as long as it provides the required methods

with this method, generating a public key 'your own way' is very easy, so the module integrates easily into your app

furthermore:

the public_key creates a captcha test / challenge ( instructions , user_prompt , correct_repsonse ) for presentation or validation
- the captcha test is handled by its own module which can be subclassed as long as it provides the required methods
- want to upgrade a test? its right there
- want a private test? create a new subclass
- want to add tests to cpan? please do!
  • the rendering is then handled by its own module which can be subclassed as long as it provides the required methods
  • the rendering doesn't just render a jpg for a visual captcha... the captcha challenge can then be rendered in any format
- image
- audio
- text

any single component can be extended or replaced - that means you can cheaply/easily/quickly create new captchas as older ones get defeated. instead of going crazy trying to make the worlds best captcha, you can just make a ton of crappy ones that are faster to make than to break :)

everything is standardized and made for modular interaction since the public_key maps to a captcha test, the same key can create an image/audio/text captcha,

Note that Render::Image is never called - it is just a base class. The module ships with Render::Img::Imager, which uses the Imager library. Its admittedly not very good- it is simple a proof-of-concept.

want gd/imagemagick? write Render::Img::GD or Render::Image::ImageMagick with the appropriate hooks (and submit to CPAN!)

This functionality exists so that you don't need to run GD on your box if you've got a mod_perl setup that aready uses Imager.

Using any of the image libraries should be a snap- just write a render class that can create an image with 'user_prompt' text, and returns 'as_string' Using any of the audio libraries will work in the same manner too.

Initial support includes the ability to have Textual logic Catptchas. They do silly things like say "What is one plus one ? (as text in english)" HTML::Email::Obfuscate makes these hard to scrape, though a better solution is needed and welcome.

One of the main points of PluggableCaptcha is that even if you create a Captcha that is one step ahead of spammers ( read: assholes ) , they're not giving up -- they're just going to take longer to break the Captcha-- and once they do, you're sweating trying to protect yourself again.

With PluggableCaptcha, it should be easier to :

a- create new captchas cheaply: make a new logic puzzle , a new way of rendering images , or change the random character builder into something that creates strings that look like words, so people can spell them easier.
b- customize existing captchas: subclass captchas from the distribution , or others people submit to CPAN. create some site specific changes on the way fonts are rendered, etc.
c- constantly change captchas ON THE FLY. mix and match render and challenge classes. the only thing that would take much work is swapping from a text to an image. but 1 line of code controls what is in the image, or how to solve it!

Under this system, ideally, people can change / adapt / update so fast , that spammers never get a break in their efforts to break captcha schemes!

new PARAMS Returns a new Authen::PluggableCaptcha object constructed according to PARAMS, where PARAMS are name/value pairs.
PARAMS are name/value pairs.

Required PARAMS are:

"type TYPE"
Type of captcha. Valid options are 'new' or 'existing'
"seed TYPE"
seed used for key management. this could be a session id, a session id + url, an empty string, or any other defined value.
"site_secret TYPE"
site_secret used for key management. this could be a shared value for your website.

Optional PARAMS are:

"keymanager_args TYPE"
The value for the keymanager_args key will be sent to the KeyManager on instantiation as 'keymanager_args'

This is useful if you need to specify a DB connection or something similar to the keymanager

"do_not_validate_key INT"
This is valid only for 'existing' type captchas.

passing this argument as the integer '1'(1) will not validate the publickey in the keymanager.

This is useful if you are externally handling the key management, and just use this package for Render + Challenge

captcha_type TYPE
get the captcha type
keymanager
returns an instance of the active keymanager
challenge_instance TYPE
returns an instance of a challenge class TYPE
render_instance TYPE
returns an instance of a render class TYPE
die_if_invalid
calls a die if the captcha is invalid
get_publickey
returns a publickey from the keymanager.
expire_publickey
instructs the keymanager to expire the publickey. on success returns 1 and sets the captcha as invalid and expired. returns 0 on failure and -1 on error.
validate_response
Validates a user response against the key/time for this captcha

returns 1 on sucess, 0 on failure, -1 on error.

render PARAMS
renders the captcha based on the kw_args submitted in PARAMS

returns the rendered captcha as a string

PARAMS are required name/value pairs. Required PARAMS are:

"challenge_class TYPE" Full name of a Authen::PluggableCaptcha::Challenge derived class
"render_class TYPE" Full name of a Authen::PluggableCaptcha::Render derived class

Set the Following envelope variables for debugging

        $ENV{'Authen::PluggableCaptcha-DEBUG_FUNCTION_NAME'}
        $ENV{'Authen::PluggableCaptcha-BENCH_RENDER'}

debug messages are sent to STDERR via the ErrorLoggingObject package

This is an initial alpha release.

There are a host of issues with it. Most are discussed here:

To Do:

        priority | task
        +++| clean up how stuff is stored / passed around / accessing defaults.  there's a lot of messy stuff with in regards to passing around default values and redundancy of vars
        +++| create a better way to make attributes shared stored and accessed
        ++ | Imager does not have facilities right now to do a 'sine warp' easily.  figure some sort of text warping for the imager module.
        ++ | Port the rendering portions of cpan gd/imagemagick captchas to Img::(GD|ImageMagick)
        ++ | Img::Imager make the default font more of a default
        ++ | Img::Imager add in support to render each letter seperately w/a different font/size
        +  | Img::Imager better handle as_string/save + support for png format etc
        -  | is there a way to make the default font more cross platform?
        -- | add a sound plugin ( text-logic might render that a trivial enhancement depending on how obfuscation treats display )

If you make your own subclasses or patches, please keep this information in mind:

        The '.' and '..' prefixes are reserved namespaces ( ie: $self->{'.Attributes'} , $self->{'..Errors'} )
        
        Generally: '.' prefixes a shared or inherited trait ; '..' prefixes an class private variable
        
        If you see a function with _ in the code, its undocumented and unsupported.  Only write code against regular looking functions.  Never write code against _ or __ functions.  Never.

Many ideas , most notably the approach to creating layered images, came from PyCaptcha , http://svn.navi.cx/misc/trunk/pycaptcha/

Jonathan Vanasco , cpan@2xlp.com

Patches, support, features, additional etc

        Kjetil Kjernsmo, kjetilk@cpan.org

Copyright (C) 2006 by Jonathan Vanasco

This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

Hey! The above document had some coding errors, which are explained below:
Around line 294:
You forgot a '=back' before '=head1'
Around line 545:
=pod directives shouldn't be over one line long! Ignoring all 2 lines of content
Around line 587:
=pod directives shouldn't be over one line long! Ignoring all 2 lines of content
Around line 657:
=pod directives shouldn't be over one line long! Ignoring all 2 lines of content
Around line 671:
=pod directives shouldn't be over one line long! Ignoring all 2 lines of content
Around line 688:
=pod directives shouldn't be over one line long! Ignoring all 2 lines of content
Around line 760:
=pod directives shouldn't be over one line long! Ignoring all 2 lines of content
Around line 822:
=pod directives shouldn't be over one line long! Ignoring all 2 lines of content
2007-02-27 perl v5.32.1

Search for    or go to Top of page |  Section 3 |  Main Index

Powered by GSP Visit the GSP FreeBSD Man Page Interface.
Output converted with ManDoc.