This application provides an environment with which to configure and
test authentication stores for the "New" authentication API ( versions
>= 0.1).

At present, there is a config file for the Minimal store, and for the
DBIx::Class store.  To use either of these make a symbolic link of the
config file you want to test to myauthapp.conf and then run prove -l
t.  As far as I know the test coverage is complete.

The authentication/authorization database schema is based on the
schema in Catalyst::Authentication::Store::DBIx::Class.

TODO
====

* Write config files for more auth stores (especially OPENID)
* Write a script to $schema->deploy to another DSN (or even an sql
  file of the right syntax)


OPENID Notes
============

some IRC transcript:

07:36 < shadowpaste> "marcus" at 212.251.165.156 pasted "how I did openid with 
                     new style auth, in case anyone else needs it" (45 lines) 
                     at http://scsys.co.uk:8001/14905

Paste ---> 
sub openid : Global {
    my ($self, $c ) = @_;
    eval {
        if ($c->authenticate({},'openid')) {
            if (my $user=$c->find_user({openid=>$c->user->{url},'dbic'})) {
                $c->set_authenticated($user,'dbic');
                return $c->res->redirect($c->uri_for('/feed',$user->obj->screenname));
            }
            my $openid:Stashed=$c->user->{url};
            $c->logout;
        }
        elsif (! @{$c->error}) {
            return if $c->res->redirect;
            $c->res->redirect($c->uri_for('/login',{openid_failed=>1}));
        }
    };
    if ($@) {
        $@ =~ s/\sat\s\S+\sline\s\d+$//;
        $c->stash->{openid_alert} =  $@;
        $c->error(0);
        $c->detach('login');
    }
    
}

# config 

authentication:
    default_realm: dbic
    realms:
        dbic:
            credentail: 
                class: Password
                password_field: password
                password_type: crypt
            store:
                class: DBIx::Class
                user_class: 'DBIC::Person'
                id_field: screenname
        openid:
            credential:
                class: OpenID
            store:
                class: Null


# although personally I'll shoot marcus for using YAML ranther than Config::General :-\


07:39 <@marcus> the way I had to do it is the same way I did it with old auth
07:36 < shadowpaste> "marcus"  [see above]
07:37 <@marcus> I just nopasted my code tho
07:38 <@marcus> so feel free to hack that into newstyle auth
07:39 <@marcus> the way I had to do it is the same way I did it with old auth
07:40 <@marcus> auth against a null store, and then manually check DBIC and set 
                the apropriate user as authenticated
07:40 <@kd> so is sub openid : Global { is in Controller::Root?
07:41 <@marcus> Controller::User in my site
07:41 <@marcus> but doesn't really matter
07:42 <@marcus> the way this works is if you manage to auth, and you have the 
                user in dbic, we set you as authenticated and redirect
07:42 <@marcus> if you manage to auth and not, we have to present a template to 
                let you tie openid to a local user or register a user
07:42 <@marcus> if you can't auth, we redirect you back to login
07:43 <@kd> sounds like a separate example then
07:44 <@marcus> well, using openid without a local dbic user seems far fetched 
                to me for 99% of the use cases
07:44 <@kd> yep
07:44 <@kd> maybe I'll paste this transcript into the README and cal it 90% done
07:44  * kd has test coverage :D
07:44 <@marcus> :D
07:45 <@marcus> when I found the null store, things went better
07:45 <@marcus> it should probably be included in the synopsis for 
                Credential::OpenID
07:46 <@marcus> but of course, that module is not hosted in our trunk :/
07:46 <@kd> yes, quite
07:47 <@kd> can we get someone who speaks japanese to persuade the developer to 
            take a cat repo commit bit?
07:47 <@kd> seen miyagawa?
07:47 <@purl> miyagawa was last seen on #catalyst 7 days ago, saying: less 
              scary than awstats anyway ;)  [Apr 24 01:14:29 2008]
07:47 <@marcus> the first part of the synopsis doesnt declare at store at all, 
                andthe one further down specifies a bogus "OpenID" store
07:44 <@kd> yep