[
https://issues.apache.org/jira/browse/TUSCANY-1679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12570742#action_12570742
]
Scott Kurz commented on TUSCANY-1679:
-------------------------------------
The implications of some other discussions we've had on the list suggest that
it would be the binding's job to switch classloaders in a case like the one I
had in mind. That is, if the client uses CL 1 and the service impl uses CL 2,
the binding should not rely on the PBVInterceptor to switch the data into the
correct CL.
So I take back stating that as a problem.
I'll see if I can rename the JIRA to focus on the need to copy faults.
> PBVInvoker always uses service-side classloader to deserialize ... also
> doesn't handle checked/business excs
> ------------------------------------------------------------------------------------------------------------
>
> Key: TUSCANY-1679
> URL: https://issues.apache.org/jira/browse/TUSCANY-1679
> Project: Tuscany
> Issue Type: Bug
> Components: Java SCA Data Binding Runtime
> Affects Versions: Java-SCA-1.0
> Reporter: Scott Kurz
> Assignee: Raymond Feng
> Fix For: Java-SCA-Next
>
>
> First, the easier issue: PBVInvoker doesn't handle faults. For a
> checked/business exc on a remotable intf, wouldn't we want the exc to get
> copied too? We wouldn't want some piece of data making its way into the
> exception to end up unexpectedly modified, right?
> Second, I have the problem that the service-side classloader is used to do
> the copy (e.g.. to do the deserialize in JavaBeansDataBinding.copy() ).
> While this is what I want when copying the inputs, it is not necessarily what
> I want when copying the outputs (or the faults which I want to copy above).
> I might, for example, want to use a client-side classloader to deserialize
> the copied objects into. Maybe it wouldn't matter if I was going to do a
> data transform anyway, but if my outputs/faults are going to go back to the
> client untransformed, then I'm going to get a ClassCastExc if they're not in
> the client classloader.
> I'm not sure how to fix this...
> Note this JIRA, as I could imagine the fixes to these two issues intersecting:
> https://issues.apache.org/jira/browse/TUSCANY-1678
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
