It's a "bug".  As you note the receiver should just ignore the disposition
and logically a peer should be able to handle a disposition for a delivery
which it has already settled without barfing (since its partner can at any
time send a spontaneous disposition update which may cross on the wire with
the settling of the receiver).

I don't think there's a JIRA already, so if you raise one with the protocol
trace attached I'll have a look at "fixing" this before the next release.

Cheers,
Rob

On 16 May 2016 at 17:57, Gordon Sim <[email protected]> wrote:

> One other thing I noticed with the latest java broker, is that it seems to
> send dispositions for deliveries that were already settled by the peer.
> E.g. for a receiver, the broker sends a transfer, the receiver responds
> with a disposition with settled=true and then the broker sends back a
> disposition for that same delivery id (also with settled=true).
>
> Though I can't find anything in the spec that directly prohibits this, it
> seems wrong to me. As the receiver will already have deleted any record of
> the delivery, it is at best redundant. Is it deliberate or a bug?
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
>

Reply via email to