Logged as: http://www.osoa.org/jira/browse/BINDINGS-159
Note that I took the liberty of changing the title of the issue to
reflect the normative statement under consideration: BJM60001.
-Eric.
On 4/11/11 8:37 AM, Simon Holdsworth wrote:
OFD0F48E78.BCCAE60D-ON8025786F.0053E71E-8025786F.0055CC5D@uk.ibm.com"
type="cite">Target:
sca-jmsbinding-1.1-csd05.pdf
Description:
The JMS binding specification
includes
statement BJM60001 with a SHOULD keyword:
For an SCA reference with a JMS
binding and
unidirectional interface, when a request message is sent as part
of a one-way
MEP, the SCA runtime SHOULD NOT set the JMSReplyTo
destination
header in the JMS message that it creates, regardless of whether
the JMS
binding has a response element with a destination
defined [BJM60001]
We need to clarify whether this
is an
optional part of the JMS binding spec.
Proposal:
SHOULD was deliberately selected
for
this statement when the JMS message exchange patterns was
reviewed and
approved. The concern expressed at the time was that for some
reason
of interoperation with an existing JMS application a one-way
interaction
may need to include a JMSReplyTo destination. Although the
interaction
is modelled as a one-way operation with no callback in SCA, a
JMS application
might still expect the JMSReplyTo destination to be set to
identify a 3rd-party
to which an additional message is to be sent. For that reason
we
did not want to preclude the JMSReplyTo from being set by the
SCA runtime
through other configuration/runtime action by making this
statement a MUST.
However we wanted to promote the practice of a null JMSReplyTo
for
one-way operations. The only way we could allow the flexibility
without
a SHOULD is to make this statement non-normative by replacing it
with something
like the following text:
The JMS specification provides
the JMSReplyTo
header as the way for a JMS application to identify the
destination on
which replies or other messages are to be placed that relate to
the one
being sent. For one-way requests sent by SCA references with
unidirectional
interfaces, the JMSReplyTo will not usually be set as no reply
or other
related message is expected.
My expectation is that setting a
non-null
JMSReplyTo for a one-way request with no callback to identify a
third party
is not a common pattern, and not one we would want to encourage
in an SOA
framework, so I would not object to a resolution which replaces
the SHOULD
by a MUST, but I wanted to air the original discussion so we
don't jump
to this conclusion.
---------------------------------
Simon Holdsworth
Unless stated otherwise
above:
IBM United Kingdom Limited - Registered in England and Wales
with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth,
Hampshire PO6
3AU
|