Ant,
In the interest of not letting the patch go stale, my first thought is to go
ahead and apply it. It passes all current test cases and samples in the
build, and if not explicitly activated, it is not likely to break things.
That being said, the standard disclaimer applies that this is largely
untested code. In addition, we need to figure out how we want to deal with
the single method in the callback interface assumption. Fixing the code to
relax it is not really the big deal, but it's not clear there is a general
approach that works across bindings. For instance, for ws/axis2 we can try
to elicit info from the wsdl at hand to associate callback methods with
forward methods. Say a given port type has two input/output operations, op1
and op2, intended to be used with callbacks. The corresponding Java
interfaces could have op1 and op2, and op1Callback and op2Callback,
respectively. The Axis2Reference code could create a callback target invoker
for op1Callback that is registered with the operation client for op1, and
similarly for op2. But notice that this is only possible because we are
using the wsdl as a hint. If, in general, all we are given are the two Java
interfaces, there is no such hint to go by.
In any case, I would also hope we can resolve this issue after the patch has
been committed. That is, assuming we can live with the simplifying
assumption initially.
----- Original Message -----
From: "ant elder" <[EMAIL PROTECTED]>
To: <[email protected]>
Sent: Thursday, September 14, 2006 5:12 AM
Subject: Re: [jira] Updated: (TUSCANY-721) Adding support for callbacks over
bindings.axis2
Ignacio, I'd like to apply this patch now, but as you said its not quite
ready to be applied I'd like your ok first.
Its going to be easier for you to continue to develop it once committed,
and
as there's a few of us all trying to work on the Axis2 binding code its
going to be impossible to stop things going stale with this just sitting
as
a patch in a JIRA. Its also much easier for people to review whats going
on
with async with it part of the trunk code. I much prefer code be developed
openly with small commits going to trunk instead of large bits of function
being completed else where and only committed once finished and polished,
so
I'm fine with this not being perfect yet and not having things like
complete
testing.
...ant
On 9/13/06, Ignacio Silva-Lepe (JIRA) <[email protected]> wrote:
[ http://issues.apache.org/jira/browse/TUSCANY-721?page=all ]
Ignacio Silva-Lepe updated TUSCANY-721:
---------------------------------------
Attachment: Axis2Callbacks.patch
This patch contains the main code and preliminary test cases added to the
existing ones. More testing, including integration testing with data
binding
and across client and server is needed, which depends on Raymond's work
on
data binding integration. So, this patch is not quite ready to be
committed.
It is being posted to allow Raymond (and anyone else interested) to
perform
more testing. It is also posted in the interest of not letting it become
stale as the work on data binding integration proceeds.
The code in this patch makes the (rather major) assumption that there is
only one callback method. Relaxing this assumption, however, does not
seem
to be trivial, it may depend on knowledge of what actual callback method
was
invoked by the service at the other end.
In addition, the assumption is made for now that the transport mep is
synchronous. That is, options.isUseSeparateListener remains false on the
operation client. If/when there is an architected way to change this
option,
the assumption can be relaxed.
> Adding support for callbacks over bindings.axis2
> ------------------------------------------------
>
> Key: TUSCANY-721
> URL: http://issues.apache.org/jira/browse/TUSCANY-721
> Project: Tuscany
> Issue Type: New Feature
> Components: Java SCA Axis Binding
> Affects Versions: Java-M2
> Reporter: Ignacio Silva-Lepe
> Assigned To: Ignacio Silva-Lepe
> Fix For: Java-M2
>
> Attachments: Axis2Callbacks.patch
>
>
> Support for callbacks over Axis2 binding is added.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
