Hi Pete,

I see that now. Stuart's idea of is much better than cloning the invocation context, as it will guarantee that the whole chain is processed correctly. Cloning the invocation context will still result in issues with decorators, in the least I can tell that for Weld.


Peter Royle wrote:
Hi Marius,

I have seen the patch and thorough explanation you put on the JIRA, thanks a lot for that. I am also talking with Stuart Douglas on IRC about a possible portable solution using ThreadLocal. I will refer probably try both ways and see which suits best just to be sure. Will report how it goes.


Pete R

On 16/05/2011, at 4:46 PM, Marius Bogoevici wrote:

Hi Pete,

NP - I'll try to put up a patch today (Monday  EDT). It's going to be a bit cumbersome for the current versions of Weld, as we need to use reflection - unfortunately the current implementation of jboss-interceptors simply doesn't account for this use case, but at least we can get something working. 

Hopefully, future versions of Weld will be able to support this in a much smoother fashion.

From: "Peter Royle" <howardmoon@screamingcoder.com>
To: "Marius Bogoevici" <mariusb@redhat.com>
Cc: "Seam-Dev" <seam-dev@lists.jboss.org>, "Weld-Dev" <weld-dev@lists.jboss.org>
Sent: Monday, May 16, 2011 2:15:49 AM
Subject: Re: [weld-dev] WELD-862 and Seam Cron

Hi Marius,

Thanks, that's hugely helpful. I hadn't realised that my implementation was outside the Interceptors spec. I guess in that case the ideal solution is one which works within the bounds of the spec (ie: the deep-copy "workaround"). I will try that out as soon as you are able to provide the patch/sample for how to do so (I had a dig around but couldn't work it out unfortunately). If I notice any significant performance implications with that approach I'll report back here for further advice.


Pete R.

On 14/05/2011, at 2:31 AM, Marius Bogoevici wrote:


Creating a separate thread from an @AroundInvoke interceptor and using that to proceed() is not actually supported by the Interceptors spec. 

"Around-invoke methods run in the same Java call stack as the associated target method."  (page 6) - which is clearly not the case here

Nevertheless, I'll try to improve the handling in jboss-interceptors (which is used by weld) to handle this use case properly. The invocationcontext is not serializable, but it can be cloned using specific stuff from jboss-interceptors, so we can activate that as a workaround for extant versions of JBoss/Glasssfish/Weld. I'll send a patch on how to do that on the issue itself.

Hope this helps,

From: "Peter Royle" <howardmoon@screamingcoder.com>
To: "Weld-Dev" <weld-dev@lists.jboss.org>
Cc: "Seam-Dev" <seam-dev@lists.jboss.org>
Sent: Friday, May 13, 2011 8:22:54 AM
Subject: [weld-dev] WELD-862 and Seam Cron


I'm aiming to make a release of Seam Cron available within the next two weeks. Currently there is an outstanding issue (https://issues.jboss.org/browse/WELD-862) which prevents Cron from running properly with Weld. I have been able to carry on developing Cron by testing it against OpenWebBeans, but obviously if we are to release a Seam module it should work against Weld.

It would be nice if WELD-862 could be fixed as soon a possible so that all future versions will work well with Cron. 

But more importantly I also probably need to do something special in Cron so that it will work with the version of Weld already deployed in JBoss AS and Glassfish, which will contain the bug. The workaround mentioned in the bug report is to deep copy the InvocationContext. I attempted to do this by serialising and unserialising the InvocationContext but couldn't due to UnserializableExceptions. Does anyone have any advice for me about how I might be able to work around this bug to support existing versions of Weld?


Pete R
weld-dev mailing list

_______________________________________________ weld-dev mailing list weld-dev@lists.jboss.org https://lists.jboss.org/mailman/listinfo/weld-dev