nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Delete callbacks in the new world

Jeff Lang

I'm trying to find a way to get a delete notification when nodes in CS are deleted. With the new method of setting the Deleted flag in DTreeCore, the old DAPICallback delete code was commented out. The current deleteInfo and deletePre nodecallback methods do run, but only for the top level node. If a container is being deleted, I need to know all the nodes within that container that are being deleted. If that container, and all its contents, are then resurrected from the delete status, how do I get notified of that operation?

Find more posts tagged with

Comments

David Templeton

Jeff

In order to get performance for the logical delete and restore we are not iterating the objects at all and no code should do so either. The idea is that on delete no solution should do anything so that on restore the world is put back transparently. So delete is not physical delete, it is logical delete. The iteration of the tree to delete each object and the subsequent metadata is done in the purge.

There are a couple of solutions then need to disconnect external resources from their tables but those are handled with deleted flags as well mirroring the DTc table. Can get into details if need there.

As for the restore callbacks there are some in the nodecallbacks already

publicfunctionAssocCBRestorePost( Object dapiCtx, DAPINODE node,Assoc restoreInfo,Dynamic context )

Assoc rtnVal

Dynamic apiError

String errMsg

Boolean ok = TRUE

rtnVal.ok = ok

rtnVal.errMsg = errMsg

rtnVal.apiError = apiError

return rtnVal

end

publicfunctionAssoc CBRestorePre( DAPINODE node,Dynamic restoreInfo = UNDEFINED,Dynamic context = UNDEFINED )

Assoc rtnVal

Dynamic apiError

String errMsg

Boolean ok = TRUE

rtnVal.ok = ok

rtnVal.errMsg = errMsg

rtnVal.apiError = apiError

return rtnVal

end

David

Jeff Lang

David,

My problem is that I'm trying to link CS to an external system. Whenever a node change is made in CS, for a specific list of node subtypes, I post a message to the external system which uses the REST API to contact CS to get information about the node and update it in that system. I'm using the dbConnect transaction callback to only post the messages when the transaction commits. My issue is that if a container is restored, I may need to send CREATE messages for all children of the container. I'm not sure if the external application knows how to rebuild the tree if this happens, so I thought I'd ask here if there was a way to get a list of all the nodes that were marked for delete or restored.

David Templeton

Well first if the restore is done there is no tree to recreate. Since the recycle bin rules have been redone completely. So any assumption about the way the old RB or ud worked is moot :-). Logical deletes are in place. Along with other rules.

You can get the list of id's if you want. But it could be in the 100k's of id's. This is the reason for the change of logic.

Select dataid from dtreeancestors where ancestorid = {parentid}

(sql not in correct case sorry :-( )

Thanks
David

Sent from my BlackBerry, expect it to be pithy.

From: eLink Entry: Content Server Development Forum

Sent: Monday, June 20, 2016 6:05 PM

To: eLink Recipient

Reply To: NjMwODcyMTYA@elinkkc.opentext.com

Subject: RE Delete callbacks in the new world

RE Delete callbacks in the new world

Posted by Lang, Jeff On 06/20/2016 06:04 PM

David,My problem is that I'm trying to link CS to an external system. Whenever a node change is made in CS, for a specific list of node subtypes, I post a message to the external system which uses the REST API to contact CS to get information about the node and update it in that system. I'm using the dbConnect transaction callback to only post the messages when the transaction commits. My issue is that if a container is restored, I may need to send CREATE messages for all children of the container. I'm not sure if the external application knows how to rebuild the tree if this happens, so I thought I'd ask here if there was a way to get a list of all the nodes that were marked for delete or restored.

[To post a comment, use the normal reply function]
Topic:	Delete callbacks in the new world
Forum:	Content Server Development Forum
Content Server:	Knowledge Center

Drane_Matthew

David,

You mentioned that there is now no iteration through the objects on logical delete.

How is permission checking accomplished? I assume it still must do a check somehow to ensure that it has rights to deleteevery object below the deleted object.

Also, if an object is on RM Hold, does it allow the logical delete (and then only block the physical delete at time or purge), or does it block the logical delete as well?

Matt

From: eLink Entry: Content Server Development Forum [mailto:development@elinkkc.opentext.com]
Sent: Monday, June 20, 2016 4:45 PM
To: eLink Recipient <devnull@elinkkc.opentext.com>
Subject: RE Delete callbacks in the new world

RE Delete callbacks in the new world

Posted by Templeton, David On 06/20/2016 04:44 PM

Jeff

As for the restore callbacks there are some in the nodecallbacks already

publicfunctionAssocCBRestorePost( Object dapiCtx, DAPINODE node,Assoc restoreInfo,Dynamic context )

Assoc rtnVal

Dynamic apiError

String errMsg

Boolean ok = TRUE

rtnVal.ok = ok

rtnVal.errMsg = errMsg

rtnVal.apiError = apiError

return rtnVal

end

publicfunctionAssoc CBRestorePre( DAPINODE node,Dynamic restoreInfo = UNDEFINED,Dynamic context = UNDEFINED )

Assoc rtnVal

Dynamic apiError

String errMsg

Boolean ok = TRUE

rtnVal.ok = ok

rtnVal.errMsg = errMsg

rtnVal.apiError = apiError

return rtnVal

end

David

[To post a comment, use the normal reply function]
Topic:	Delete callbacks in the new world
Forum:	Content Server Development Forum
Content Server:	Knowledge Center

David Templeton

Matt

How is permission checking accomplished? I assume it still must do a check somehow to ensure that it has rights to deleteevery object below the deleted object.

Yeap it is checked for all objects.

Also, if an object is on RM Hold, does it allow the logical delete (and then only block the physical delete at time or purge), or does it block the logical delete as well?

RM hold is done in set wise operation and it does block the logical delete.

David

HFCDev

Hi David et al,

I have a related question but it has more to do with Move than Delete. We're trying to manage very large moves here at CIC. We have build a daemon that uses Web Services to perform moves in a one by one fashion so that an entire folder tree is flattened in a temp folder, then reassembled one node at a time in the target folder. This means that for a really big folder tree, the database is not getting hammered by the move (I'm not sure how the database gets hammered but it's the reason that CIC has a move throttle imposed that restricts moves to 500 items), and users wanting folders moved don't need to know anything about the size of the tree.

(now you understand why I was talking about node cralwers in another thread, but for now, we're doing this in Web Services not using Oscript)

What we're not sure how to manage is users either performing their own moves within the source tree, or doing deletes or additions while we're flattening the folder tree. Would a hold imposed on the top level folder prevent these activities but not prevent the daemon running as an admin service account?

The only alternative I see is a programmatic one where I create a node move callback that will check that the ID of the src or destination is not within the tree being moved. Given the strict no-customization policy, I'll never get that through.

-Hugh