jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alfie Kirkpatrick" <Alfie.Kirkpatr...@ioko.com>
Subject RE: Observation before node delete
Date Tue, 05 May 2009 08:41:54 GMT
This seems like a big hoop to jump through for such a 'simple'
requirement, and would require my application to fully cooperate. I'm
looking for something a bit more cross-cutting.

Have looked at wrapping Session and Node to intercept changes, eg.
Node.remove(). This is possible but there are other ways nodes can be
deleted. For example an import with
IMPORT_UUID_COLLISION_REMOVE_EXISTING might end up removing nodes which
would not be caught by this approach.

Regards, Alfie.

-----Original Message-----
From: Alessandro Bologna [mailto:alessandro.bologna@gmail.com] 
Sent: 01 May 2009 13:24
To: users@jackrabbit.apache.org
Cc: <users@jackrabbit.apache.org>
Subject: Re: Observation before node delete

We had a similar use case, and we decided to simply flag the node with  
a "remove" property, and let the listener remove it. In this way you  
can of course access the node...


On May 1, 2009, at 6:25 AM, "Alfie Kirkpatrick"
 > wrote:

> Hi, am wondering if it's possible to hook into change events <before>
> they get persisted? In particular we want to access information  
> about a
> node before it is deleted, for example its node type or property  
> values.
> In some scenarios we may even want to prevent a delete operation if
> certain criteria about the node are met/not met, for example by  
> throwing
> an exception in the event handler.
> The observation API only gives a path and, in Jackrabbit's
> implementation, even using a SynchronousEventListener the node is gone
> by the time the delete event gets fired. Are there any other hooks
> available/feasible?
> An alternative solution might be to somehow get access to the  
> session's
> change list before calling save, and then access the deleted node in
> another session. But there is no obvious place to get this information
> outside the observation framework.
> Is anyone else interested in this problem?
> Many thanks!
> Alfie.

View raw message