Use SecurityInfo throughout instead of User directly (currently inconsistent)

[blmstr]

Implemented OnBehalfOf functionality along the lines of your blog post (http://ayende.com/blog/4618/security-models-on-behalf-of) successfully but had to make a couple of changes to rhino security.

SecurityInfo isn't constantly used to get the user id (was in SecurityCriterions for example but not in these two files) so have implemented this minor change.

Allows a user to act on behalf of another by simple returning another user's SecurityInfo, auditing etc can be handled against the user that is currently logged in.

Updated to use 'user.id' instead of 'user.Id'