CanCan is an authorization solution for Ruby on Rails. This restricts what a given user is allowed to access throughout the application. It is completely decoupled from any role based implementation and focusses on keeping permission logic in a single location (the +Ability+ class) so it is not duplicated across controllers, views, and database queries.
This assumes you already have authentication (such as Authlogic[http://github.com/binarylogic/authlogic] or Devise[http://github.com/plataformatec/devise]) that provides a +current_user+ method which CanCan relies on. See {Changing Defaults}[http://wiki.github.com/ryanb/cancan/changing-defaults] if you need different behavior.
Setting this for every action can be tedious, therefore the +load_and_authorize_resource+ method is provided to automatically authorize all actions in a RESTful style resource controller. It will use a before filter to load the resource into an instance variable and authorize it for each action.
If the user authorization fails, a CanCan::AccessDenied exception will be raised. You can catch this and modify its behavior in the +ApplicationController+.
As shown above, the +Ability+ class is where all user permissions are defined. The user model is passed into the initialize method so the permissions can be modified based on any user attributes. CanCan makes no assumptions about how roles are handled in your application. See {Role Based Authorization}[http://wiki.github.com/ryanb/cancan/role-based-authorization] for an example.
The +can+ method is used to define permissions and requires two arguments. The first one is the action you're setting the permission for, the second one is the class of object you're setting it on.
You can pass an array for either of these parameters to match any one. In this case the user will have the ability to update or destroy both articles and comments.
can :manage, Article # has permissions to do anything to articles
can :read, :all # has permission to read any model
can :manage, :all # has permission to do anything to any model
You can pass a hash of conditions as the third argument to further restrict what the user is able to access. Here the user will only have permission to read active projects which he owns.
If the block returns true then the user has that :+update+ ability for that project, otherwise he will be denied access. See {Defining Abilities with Blocks}[http://wiki.github.com/ryanb/cancan/defining-abilities-with-blocks] for more information.
You will usually be working with four actions when defining and checking permissions: :+read+, :+create+, :+update+, :+destroy+. These aren't the same as the 7 RESTful actions in Rails. CanCan automatically adds some default aliases for mapping those actions.
Notice the +edit+ action is aliased to +update+. If the user is able to update a record he also has permission to edit it. You can define your own aliases in the +Ability+ class
In the controller +index+ action you may want to fetch only the records which the user has permission to read. You can do this with the +accessible_by+ scope.
CanCan was inspired by declarative_authorization[http://github.com/stffn/declarative_authorization/] and aegis[http://github.com/makandra/aegis]. Many thanks to the authors and contributors.