Folks,
I'm not keen on double negatives, so
let me try some wordsmithing of the second part:
A class which provides a service with
a conversational interface can have any scope. In particular, it
is not necessary for the class
to have conversation scope. If
the class has conversation scope, the class benefits from the runtime maintaining
state associated
with the conversation, typically through
routing each operation invocation associated with the conversation to the
same
instance of the class, with state data
held in instance variables within the class. However, for classes
of any scope, when an
operation of a conversational interface
is executing, the ComponentContext.getConversationID() returns the conversation
ID of
the conversation. The conversation
ID can be used by the class as an index to store and to look up state data
associated with the
conversation, using some suitable storage
mechanism.
OK, this is a longer set of words, but
I hope it explains things more clearly by being explicit about what is
going on.
Yours, Mike.
Strategist - Emerging Technologies, SCA & SDO.
Co Chair OASIS SCA Assembly TC.
IBM Hursley Park, Mail Point 146, Winchester, SO21 2JN, Great Britain.
Phone & FAX: +44-1962-818014 Mobile: +44-7802-467431
Email:
"Michael Rowley"
<>
07/02/2008 19:33
To
"Peshev, Peter" <>,
Simon Nash/UK/IBM@IBMGB, <>
cc
Subject
RE: [sca-j] ISSUE 12: Relation between
conversational annotation and scope conversation
On today’s call, we resolved
the first part of this issue as follows:
Resolved: Add these words to
section 2.4.4: "A conversational scoped class MUST NOT expose a service
using a non-conversational interface."
We discussed possible wording
for the second part of the proposal. I made some attempts at wording
during the meeting that weren’t very good. Here is my latest attempt:
When a class provides a service
with a conversational interface, it does not implicitly affect the scope
of that class. If the scope is not conversational, then the runtime
will not automatically maintain state that is associated with the conversation.
However, when an operation of a conversation interface is executing,
the ComponentContext.getConversationID() will return the conversation ID
for the conversation, irrespective of the scope of the component.
Michael
From: Peshev, Peter [mailto:]
Sent: Thursday, February 07, 2008 1:03 PM
To: Simon Nash;
Subject: RE: [sca-j] ISSUE 12: Relation between conversational annotation
and scope conversation
Hi Simon,
I am Ok with that suggestion.
I am just trying to clarify some
parts of the spec which are left not clearly stated.
NEW PROPOSAL :
It should be clarified that in the above example -
Having a StatefullClass exposing service over a non-conversational interface
MUST generate an error at deployment.
The instances of the class StatelessClass are NOT REQUIRED to contain
conversational state between methods; Any instance can be used for any
client. An SCA runtime MAY provide a
pool of such objects and reuse them among calling clients or MAY instantiate
each time a new instance. ConversationalId MUST be provided
since the interface was marked via @Conversational
From: Simon Nash [mailto:]
Sent: Thursday, 7. February 2008 19:37
To:
Subject: RE: [sca-j] ISSUE 12: Relation between conversational annotation
and scope conversation
I'm OK with the second part of the proposal, but not with the first part.
When a conversation is initiated, action is needed on the client
side as well as the service side. If the interface is not marked
conversational, the client can't know that it should perform the necessary
actions. These include creating a conversation ID and sending it
on the wire.
I think this combination of a non-conversational interface and a conversational
implementation should be treated as an error.
Simon
Simon C. Nash, IBM Distinguished Engineer
Member of the IBM Academy of Technology
Tel. +44-1962-815156 Fax +44-1962-818999
"Peshev, Peter"
<>
05/10/2007 09:20
To
<>
cc
Subject
RE: [sca-j] ISSUE 12: Relation between
conversational annotation and scope conversation
Resending the email in this mail thread, just to keep track :
DESCRIPTION:
The annotation @Conversational is put on interfaces, while the
annotation @Scope is put on the implementation types for POJO-s. They
are used to control whether the runtime will keep state or not between
invocations. (Some other future implementation types in scope of
the TC
may provide their own stateless and stateful semantics equivalent to the
scope mechanics)
It should be clarified what should happen if @Conversational is put on
a
service interface however the class which is exposing that service is
lacking @Scope(Conversation)
It should be clarified what should happen if there is NO @Conversational
on a service interface however the class which is exposing that service
is having @Scope(Conversation).
More precisely having in mind the following definitions (each in its own
file) :
import org.osoa.sca.annotations.Conversational;
@Conversational public interface ConversationalInterface {}
---
public interface NonConversationalInterface {public void test();}
---
import org.osoa.sca.annotations.Scope;
import org.osoa.sca.annotations.Service;
@Service(NonConversationalInterface.class)@Scope("CONVERSATION")
public class StatefullClass implements NonConversationalInterface {
public void test() {}}
---
import org.osoa.sca.annotations.Service;
@Service(ConversationalInterface.class)
public class StatelessClass implements ConversationalInterface{
public
void test() {}}
It should be clraified whether it is legal to deploy such components and
what should happen if a client calls via sca two times the method test()
for ServiceReferences correspondant to the two classes. Is the runtime
responsible to keep the state between the invocations ?
PROPOSAL :
It should be clarified that in the above example -
Instances of the StatefullClass MUST contain conversational state which
MUST be retained across methods and transactions called from the client.
The instances of the class StatelessClass are NOT REQUIRED to contain
conversational state between methods;
Any instance can be used for any client. An SCA runtime MAY provide a
pool of such objects and reuse them among calling clients or MAY
instantiate each time a new instance. ConversationalId will be provided
since the interface was marked via @Conversational
-----Original Message-----
From: Barack, Ron [mailto:]
Sent: Thursday, 4. October 2007 21:07
To:
Subject: [sca-j] ISSUE 12: Relation between conversational annotation and
scope conversation
http://www.osoa.org/jira/browse/JAVA-12
-----Ursprüngliche Nachricht-----
Von: Peshev, Peter [mailto:]
Gesendet: Donnerstag, 27. September 2007 17:48
An:
Betreff: [sca-j] NEW ISSUE: Relation between conversational annotation
and scope conversation
TARGET:
Java Common Annotations and APIs specification
DESCRIPTION:
The specification mentions @Conversational and @Scope(Conversation) but
there is no clarification how these interfere together and what should
happen in all the possible combinations.
PROPOSAL:
Two new paragraphs should be added in section (3.2 @Conversational) that
have the following wording :
In case an interface is marked as conversational but the scope of
the
target implementation is different than @Scope(Conversation), than the
SCA runtime would invoke an instance as defined by the scope. In case
the scope is the default one (stateless) than the container may dispatch
to a new instance each time or alternatively pull one from a pool. In
this case, it is assumed that the implementation itself will manage
state. The implementation would be responsible for using the
conversation id associated with the request (obtaining it through
injection or via the SCA API) to obtain state stored somewhere (cache ,
database , etc.).
In case the target implementation has @Scope(Conversation) and the
interface is NOT marked as conversational than there will be no
conversation, attempts to retrieve conversationId will result in null,
and the SCA runtime may behave as if for that particular invocation the
scope has been defined as stateless.
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
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
