Well, seems this first part was not hard at all.
I've attached a patch on https://issues.apache.org/jira/browse/OPENEJB-1165
Included the AsynchronousRunner and AsynchronousRunnerTest.

Luis Fernando Planella Gonzalez


Em Quinta-feira 21 Janeiro 2010, às 15:14:00, David Blevins escreveu:
> 
> On Jan 21, 2010, at 4:45 PM, Luis Fernando Planella Gonzalez wrote:
> 
> >> Help is always welcome.  This one is tricky to describe, so I'm going
> >> to try the Mr. Wizzard approach and list the "ingredients" first.   
> >> Ok,
> >> so for this functionality, we need:
> >>
> >>  - a thread pool (java.util.concurrent.Executor) -- can use a
> >> standard impl
> >>  - a java.util.concurrent.FutureTask
> >>  - a java.util.concurrent.Callable -- we will have to make one of  
> >> these
> >>
> >> The callable will be a class that does the actual execution of the
> >> async method, we'll be constructing one per method call passing in  
> >> the
> >> method to be invoked, the object itself, and the parameters.
> >>
> >> In the main request thread we will not invoke the target method
> >> directly, instead we will:
> >>
> >>   1. construct a Callable object that contains all the information to
> >> do the invocation
> >>   2. construct a FutureTask using the Callable
> >>   3. give the FutureTask to the Executor
> >>   4. return the FutureTask as the return value
> >>
> >> We don't need to plum all this into the container code just yet.
> >> Initially we just need a little class (maybe add it to
> >> org.apache.openejb.util) that wraps all of this up and some test  
> >> cases
> >> for it.  We can worry about the when and where to use it question
> >> later -- it's much tricker and far less isolated.
> >>
> >> Hopefully this is enough to get started if you're interested in
> >> hacking on it.
> > Ok, leave it to me, I'll hack on this, as it seems interesting and  
> > not extremely hard...
> > However, if things go wrong, I'll come back crying out for help ;)
> 
> Cool.  Feel free to chat on the list conversationally while coding.   
> Never hurts.
> 
> >>> And, do you have an idea about when OpenEJB 3.1.3 will be released?
> >>
> >> No date has been set yet, do you have a preference?
> > No, just want to be sure that it will be released, and not just be  
> > forgotten because of the version 3.? / 4.? which will implement EJB  
> > 3.1 xD
> 
> So far we've been adding EJB 3.1 features in to the 3.1.x branch.   
> There are only a few parts of EJB 3.1 that involve a change to  
> existing ejb interfaces, so hopefully we can get pretty towards 3.1  
> before we feel the need to switch version numbers.
> 
> 
> 

Reply via email to