This is now logged as
issue i092.
From: Doug Davis
[mailto:dug@us.ibm.com]
Sent: Thursday, January 26, 2006
2:46 PM
To: ws-rx@lists.oasis-open.org
Subject: Re: [ws-rx] NEW ISSUE:
Where is a SequenceAcknowledgement sent on receipt of a AckRequested header?
Anish,
If you could write up some replacement text for those offending "returned/response"
words I think you're change would be ok. Tweaking the existing text
instead of adding more text would be best.
thanks,
-Doug
Anish Karmarkar
<Anish.Karmarkar@oracle.com>
01/26/2006 04:10 PM
|
To
|
Christopher B Ferris/Waltham/IBM@IBMUS
|
cc
|
wsrx <ws-rx@lists.oasis-open.org>
|
Subject
|
Re: [ws-rx] NEW ISSUE: Where is a
SequenceAcknowledgement sent on receipt of a AckRequested header?
|
|
Yes,
I did see that statement, but is was not clear to me that it
applied to every SeqAck header that is sent by the
RMD. There are three
cases when SeqAcks are sent:
1) when the RMD decides to send it unilaterally.
2) when the RMD decides to send it on receipt of a
AckRequested header
3) when the RMD receives a CloseSequence message
and sends a
CloseSequenceResponse message along with the
SeqAck header.
In the 3rd case, it is required that the SeqAck
header go back to the
replyTo address (along with the
CloseSequenceResponse message). In the
1st case, it is clear that the SeqAck is sent to
AcksTo EPR. The 2nd
case is ambiguous to me.
The spec in section 3.5 says:
"The purpose of the <wsrm:AckRequested>
header block is to signal to the
RM Destination that the RM Source is requesting
that a
<wsrm:SequenceAcknowledgement> be
****returned****.
The RM Source may request an acknowledgement
message from the RM
Destination at any time by including an
<wsrm:AckRequested> header block
in the message. An RM Destination that receives a
message that contains
an <wsrm:AckRequested> header block MUST
****respond**** with a message
containing a <wsrm:SequenceAcknowledgement>
header block."
It wasn't clear to me that the stmt that you point
out applied in the
2nd case. Perhaps it was just me. If so, obviously
it isn't an issue.
> Making such a statement redundantly elsewhere
> in the spec seems unnecessary, IMO.
We don't have to make a redundant statement
elsewhere. We can modify the
existing statement (that you pointed out) in place
to make this clear.
-Anish
--
Christopher B Ferris wrote:
>
> CDII draft states on line 258:
>
> /wsrm:CreateSequence/wsrm:AcksTo
> This REQUIRED element, of type
wsa:EndpointReferenceType as specified by
> WS-Addressing [WSAddressing]
> specifies the endpoint reference to which
<wsrm:SequenceAcknowledgement>
> messages
> and faults related to the created Sequence
are to be sent.
>
> I see no ambiguity here. Making such a
statement redundantly elsewhere
> in the spec seems unnecessary, IMO.
>
> Cheers,
>
> Christopher Ferris
> STSM, Emerging e-business Industry
Architecture
> email: chrisfer@us.ibm.com
> blog:
http://www.ibm.com/developerworks/blogs/dw_blog.jspa?blog=440
> phone: +1 508 377 9295
>
> Anish Karmarkar
<Anish.Karmarkar@oracle.com> wrote on 01/26/2006
> 02:54:27 PM:
>
> > Title: Where is the
SequenceAcknowledgement sent on receipt of
> > AckRequested header?
> > Description: The spec does not say
where the SequenceAcknowledgement
> > message is sent on receipt of the
AckRequested header. There are two
> > possibilities: to the AcksTo EPR
or to the ReplyTo (as a response) of
> > the message requesting the SequenceAcknowledgement.
> > Target: core
> >
> > -Anish
> > --
> >