Arg #TOTSCO
They casually send out a bulletin on handing incoming messages, time to respond, and handing duplicated. The duplicate handling advised basically MANDATES that senders have a unique source correlationID on messages. The spec does not say that (do tell me if I am wrong) and the examples have duplicates.
So updating NOTSCO to report duplicates now.
Why can we not have a good specification?
@revk
"no, we can't fix this obviously broken spec because it's under a change freeze"
"here's a bulletin fixing a small area of brokenness, but it's totally not breaking the change freeze because it's only a bulletin".
Yep, sounds about right.
@falken @ahnlak the bulletin does not even say senders have to be unique - it just explains receivers can treat as a duplicate message. So even the change is just “implied”.