Class IPermissionPolicy

Check that the action can be performed by username on the resource

Note that when checking a permission on a realm resource (i.e. when .id is None), this usually corresponds to some preliminary check done before making a fine-grained check on some resource. Therefore the IPermissionPolicy should be conservative and return:

• True if the action can be allowed for some resources in that realm. Later, for specific resource, the policy will be able to return True (allow), False (deny) or None (don't decide).
• None if the action can not be performed for some resources. This corresponds to situation where the policy is only interested in returning False or None on specific resources.
• False if the action can not be performed for any resource in that realm (that's a very strong decision as that will usually prevent any fine-grained check to even happen).

Note that performing permission checks on realm resources may seem redundant for now as the action name itself contains the realm, but this will probably change in the future (e.g. 'VIEW' in ...).

Parameters:

• action - the name of the permission
• username - the username string or 'anonymous' if there's no authenticated user
• resource - the resource on which the check applies. Will be None, if the check is a global one and not made on a resource in particular
• perm - the permission cache for that username and resource, which can be used for doing secondary checks on other permissions. Care must be taken to avoid recursion.

Returns:

True if action is allowed, False if action is denied, or None if indifferent. If None is returned, the next policy in the chain will be used, and so on.