I'm investigating the powerful pubsub in ejabberd 2.0 where items are persistent.
In my implementation there will be a very large number of nodes in a collection node
and a large number of items in a leaf node.
Is there any paging implemented in the service discovery in ejabbaerd?
In the pubsub protocol I saw I can only retrive all the items in a leaf node or a number of the last submitted items
but it suggest to use the "result set management"
Is it implemented or can I use something else?
Thanks
I asked the author of the
I asked the author of the recent PubSub module included in ejabberd 2.0.0.
You provided three alternatives:
A) retrive all the items in a leaf node
B) or a number of the last submitted items
C) but it suggest to use the "result set management"
A is implemented, of course.
B is technically implemented, but the calls are not included, and hence is not usable right now. That means send_items have a number parameter, but is never used. Currently pubsub uses send_all_items and send_last_item. Probably a patch could implement send_last_items(Number), that would call send_items(Host, Node, LJID, Number).
C or similar is not implemented. Yes, it can be useful. If you implement, you can publish a patch to be reviewed and included in ejabberd.