RE: [ws-rx] PR007: concrete proposal

[Matthew Lovett <MLOVETT@uk.ibm.com>]

Hi Gil,

One small issue, possibly just a typo. Most of the 'Terminating' states 
you created return uknown sequence faults, but I'd have thought we should 
return sequence terminated faults. Is that ok with you?

Thanks

Matt



"Gilbert Pilz" <gpilz@bea.com> wrote on 10/01/2007 23:48:49:

> I think you are right about removing the "Closing" state from the RMD 
table
> and moving straight to the "Closed" state. The spec says that the RMD 
has to
> include a SeqAck+Final header when it sends a CloseSequence message. 
This
> means it can't accept any more Sequence Traffic messages because that 
would
> invalidate this header. This means that, effectively, the RMD's state 
moves
> from "Created" to "Closed" when the RMD sends a CloseSequence message.
> 
> The rest sounds right to me as well. Attached is another draft of the
> proposal with a "Terminating" state added to the RMD table.
> 
> - gp
> 
> p.s. I would like to commend those that worked so hard to add the state
> tables to the spec. Without them I doubt we would be thinking things 
through
> to this level of detail . . .
>