Hi, everyone, I have a small problem with the monitoring plugin’s treatment of stale conversations.
Currently the “idleTime” parameter is honored when a new message for a given conversation is stored.
This means that if say a conversation is stale after 10 minutes of inactivity per settings, a query to the monitoring plugin using either xep136 will make it look active.
We usually ask for conversations using xep136 to list open and closed conversations, and the make queries using xep313 for the actual fetch (Since having the full stanza retrieved is advantageous to us).
This means that given that no indication is given from the protocol as to how many messages are in an open (or closed) conversation client applications in doubt have to refetch the last conversation.
This means that we are refetching messages for people who haven’t written anything in months sometimes.
I think it would be advantageous if a TimerTask (Or similar preferred approach) cleaned up stale open conversations periodically.
It would also be good if it was possible to add an optional conversation hash or counter allowing clients to discerne whether they actually need to fetch.
Thanks in advance,