jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Michael Neale" <michael.ne...@gmail.com>
Subject Re: How should a Custom AccessManager access a Node ?
Date Tue, 17 Oct 2006 19:37:04 GMT
Yes I think by "listening" for changes and updating an ACL, it can be kept
fast and in memory (at a small cost of some extra memory - but ACLs don't
come for free !).
I would be keen to see whatever you come up with.

Michael.

On 10/17/06, Sriram Narayanan <sriramnrn@gmail.com> wrote:
>
> On 10/13/06, Michael Neale <michael.neale@gmail.com> wrote:
> > Sririam - would you care to add some info to the wiki on your findings?
> > While access manager is not part of JCR, woudl be nice to have a wiki
> page
> > on how to use it effectively to create ACLs as you are doing.
> >
>
> My apologies for the delay. I've been busy with my project, and have
> not made time for JackRabbit.
>
> - I'd tried to use another session to connect to the repository and
> then lookup a Node in order to look up its attributes. I get a
> PathNotFoundException. I'm unable to retrieve Node for some reason.
> This is in case of even Nodes that are already present in the
> repository.
>
> For now, I've decided to not access the Node itself. I'm instead
> checking the node path (the hierarchyManager.getPath(itemID) gives you
> a path like "*{}Parent{}Child{}GrandChild". I'm restricting myself at
> the moment to checking the path text to decide whether the accessor
> has permissions or not.
>
> I now see the wisdom of posts by experienced JackRabbit users who
> recommend that one listen to the repository for changes, and then
> update an ACL based on these changes.
>
> I'm also reading the JCR PDF document, and hope to come up with an
> Authentication and Authorization implementation for JackRabbit that
> folks could use.
>
> I'l update the wiki tomorrow, since I'm exhausted right now.
>
> Thanks, everyone, who's responded to my requests for help.
>
> -- Sriram
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message