Hi Jens
Thanks for reporting this. I have looked into it.
Yes the seda is async by nature and should not spin off a copy of the
exchange that shares the same unit of work as the original.
I have a fix at hand and will commit it later today.
BTW The same issue applies for the wire tap. It should also be a new
copy not sharing the same unit of work.
I will create a ticket so we have it on record.
On Wed, Apr 8, 2009 at 3:21 PM, _Jens <jens.riemschnei...@icw.de> wrote:
>
> Hi,
>
> I ran into a problem with a route that called a seda endpoint. For some
> reason the unit of work within the seda route was completed before the
> actual route processing finished. Here's a simplified test case:
>
> public class SedaUnitOfWorkTest {
> private String lastOne;
>
> �...@test
> public void testSeda() throws Exception {
> CamelContext context = new DefaultCamelContext();
> context.addRoutes(new RouteBuilder() {
> �...@override
> public void configure() throws Exception {
> from("direct:start")
> .to("seda:foo")
> .delayer(2000);
>
> from("seda:foo")
> .process(new Processor() {
> �...@override
> public void process(Exchange exchange) throws
> Exception {
> exchange.getUnitOfWork().addSynchronization(new
> Synchronization() {
> �...@override
> public void onComplete(Exchange exchange) {
> lastOne = "UnitOfWork";
> }
>
> �...@override
> public void onFailure(Exchange exchange) {
> lastOne = "UnitOfWork";
> }
> });
> }
> })
> .delayer(4000)
> .process(new Processor() {
> �...@override
> public void process(Exchange exchange) throws
> Exception {
> lastOne = "Processor";
> }
> });
> }
> });
>
> context.start();
> context.createProducerTemplate().send("direct:start", new
> DefaultExchange(context));
>
> Thread.sleep(7000);
>
> assertEquals("UnitOfWork", lastOne);
> }
> }
>
> I'm expecting that the unit of work is the last thing that sets the field
> 'lastOne'. But it isn't, the processor is called after the unit of work
> finished.
>
> What happens is that the exchange within the Seda route does not have its
> own UnitOfWork, it is using that of the other route. This might be ok if the
> processing of the second route would be synchronous. But as it is not, I
> would expect that the second route either has its own unit of work or
> onComplete is only called if both routes have finished processing the
> exchange. Or am I wrong?
>
> Thanks,
> Jens
>
> --
> View this message in context:
> http://www.nabble.com/UnitOfWork-ends-too-soon-with-Seda-endpoints-tp22950359p22950359.html
> Sent from the Camel - Users (activemq) mailing list archive at Nabble.com.
>
>
--
Claus Ibsen
Apache Camel Committer
Open Source Integration: http://fusesource.com
Blog: http://davsclaus.blogspot.com/
Twitter: http://twitter.com/davsclaus
Apache Camel Reference Card:
http://refcardz.dzone.com/refcardz/enterprise-integration
