will do.

On 8/28/07, James Strachan <[EMAIL PROTECTED]> wrote:
> BTW am thinking - this patch doesn't affect too much yet - mostly the
> seda & file components along with the Pipeline - so how about we
> commit it; then we can experiment with different ways to improve it as
> we also try fix CAMEL-123?
>
>
>
> On 8/28/07, James Strachan <[EMAIL PROTECTED]> wrote:
> > On 8/28/07, Hiram Chirino <[EMAIL PROTECTED]> wrote:
> > > Got those test failures fixed now..  Here is a better version of the 
> > > patch.
> >
> > Great stuff!
> >
> > I've been thinking we need a way to register onComplete / onFail
> > hooks. (Rather like TransactionSynchronization in Spring)...
> > http://static.springframework.org/spring/docs/2.0.x/api/org/springframework/transaction/support/TransactionSynchronization.html
> >
> > So allowing something like
> >
> > exchange.getTransaction().addSynchronization(new Synchronization() {
> >   pubic void onCommit() {
> >      //  remove the file...
> >   }
> >
> >   public void onRollback() {
> >      ///
> >   }
> > }
> >
> > https://issues.apache.org/activemq/browse/CAMEL-123
> >
> > If we had some kinda Transaction object, which was propogated across
> > any copied exchanges (e.g. a new exchange for each async operation or
> > when using multicast etc), then we'd have a place we could register
> > these kinds of onCommit/onRollback handlers. Then each component in
> > the pipeline - whether file or ftp or whatever, could add their own
> > onCommit/onRollback handlers etc.
> >
> > If we had this single Transaction object which is properly propogated,
> > maybe that could also take over some of the work doing the tracking of
> > the number of async steps per transaction as well as being the
> > AsyncCallback - to contain the count down latch and so forth. So
> > moving some of the code from Pipeline into this single place, this
> > Transaction object - which hopefully could make it a bit easier to
> > handle async processors in some of the other processors with minimal
> > code etc.
> >
> > Am just wondering if we can minimise the amount of work required in
> > the pipeline/processor code to support async handling.
> >
> > James
> > -------
> > http://macstrac.blogspot.com/
> >
>
>
> --
> James
> -------
> http://macstrac.blogspot.com/
>


-- 
Regards,
Hiram

Blog: http://hiramchirino.com

Reply via email to