[ 
https://issues.apache.org/jira/browse/JCR-3451?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

angela resolved JCR-3451.
-------------------------

    Resolution: Invalid

if you remove a user/group node directly instead of using user management API, 
you are in charge
of clearing/updating the membership yourself.

i don't think that this something jackrabbit needs to deal with in the first 
place (except for making the
user/group hierarchy protected again which would need a separate discussion).
                
> Deleting a group node directly doesn't revoke cached memberships
> ----------------------------------------------------------------
>
>                 Key: JCR-3451
>                 URL: https://issues.apache.org/jira/browse/JCR-3451
>             Project: Jackrabbit Content Repository
>          Issue Type: Bug
>          Components: security
>    Affects Versions: 2.4
>            Reporter: Honwai Wong
>
> If a group node is deleted directly (e.g. beneath /home/groups) via the JCR 
> API, previously effective group memberships are not revoked and purged 
> consistently from the membership cache. This is potentially a security risk 
> where inherited permissions from indirect group-memberships will become 
> effective.
> Following the use-case:
> - create user testuser
> - create group testgroup
> - make testuser member of testgroup
> - deny read access for testgroup on /content/xyz
> - create another group parentgroup
> - allow read access for parentgroup on /content/xyz
> - make testgroup member of parentgroup
> --> at this point, testuser doesn't have read access to /content/xyz
> - delete testgroup by removing the corresponding group node in the repository 
> (e.g. /home/groups/t/testgroup)
> --> testuser has read access to /content/xyz !
> What seems to be missing is a call to 
> org.apache.jackrabbit.api.security.user.Authorizable#remove which ultimately 
> triggers the ClearMembershipAction to take care of cleaning up stale group 
> memberships. A potential solution would be to implement an observation based 
> approach to catch node removals of group nodes.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Reply via email to