Access Control Layer

This development is designed to allow access restrictions to be placed on EPrints, Documents and other data objects.

The model we're currently work with is:

Authenticating Authority (AA) - a trusted sign-in method that may also provide some knowledge about the person signing in, and offer groupings based on their attributes.

ACLRole - These are configured based on what an AA can offer. When a user signs in, any ACLRoles that the AA can grant to the user are added to their session. A user could sign in via more than one AA at a time.

ACLGroup - A set of ACLRoles that can be applied to an EPrint/Document/?. When access to the item is requested, the session ACLRoles are checked against the allowed ACLRoles. If there is a match, access is granted.

When we make this shizzle, how should it be done?

• StyleGuide
• NO text in citation files or other files - use phrases - a la https://github.com/eprints/eprints/commit/cccfab14e0b6764d7f1d80daf57bf4e308f56dd7

Things to do

• LogHandler - what do we log an when?
• Check poly-lingual behaviour
• Check behaviour for summary_page and summary_page_denied
• Abstract summary page handler up to dataobj level
• Add config options to enable summary pages for specific dataobjs $c->{summary_pages}->{document} = sub{ get eprint -> is archive} type thing

Other thoughts / useful links

• Category:Authentication
• http://files.eprints.org/836/ - Webserver authentication using auth_memcookie and simplesamlphp
• Webserver authentication - Single-sign-on via Shibboleth, CAS/mod_cas, Kerberos or just about any mod_auth_* Module for Apache httpd
• CAS - Central Authentication Service
• LDAP - Lightweight Directory Authentication Protocol
• Shibboleth authentication - UK Federation
• https://www.google.co.uk/search?q=site%3Afiles.eprints.org+authentication - Google files.eprints.org for Authentication
• https://www.google.co.uk/search?q=site%3Abazaar.eprints.org+authentication - same, but on the Bazaar.