OK thanks, sounds like a plan - yep, as this has been same 10 years, changing
this in 2.x would break backward compatibility, so it's fair enough to keep as
is and revisit for 3.0 :)
https://issues.apache.org/jira/browse/CAMEL-10910
Thanks,
Tomo
On 02/28/2017 11:58 PM, Claus Ibsen wrote:
Hi
This is the behavior of 2.x and the EIPs and routing engine is
optimized / doing this.
We have already done too many changes IMHO in 2.19.x and should not do
more changes for the sake of a change.
Camel 3.0 allows to revisit the routing engine and exchange copy stuff
that happens.
You are welcome to log a JIRA for 3.0.
On Tue, Feb 28, 2017 at 3:41 PM, Tomohisa Igarashi
<tm.igara...@gmail.com> wrote:
Hi,
I noticed that if eventDrivenProcessors contains only one processor, IN
message is not copied to OUT even when exchange pattern is InOut while it
does if it contains multiple processors.
eventDrivenProcessors are wrapped with Pipeline here when initiating a camel
route here:
https://github.com/apache/camel/blob/master/camel-core/src/main/java/org/apache/camel/impl/DefaultRouteContext.java#L161
But if there's only one event driven processor, the processor itself is used
instead of wrapping with Pipeline:
https://github.com/apache/camel/blob/master/camel-core/src/main/java/org/apache/camel/processor/Pipeline.java#L57
Pipeline copies the IN message to OUT if it's InOut and OUT message doesn't
exist:
https://github.com/apache/camel/blob/master/camel-core/src/main/java/org/apache/camel/processor/Pipeline.java#L107
So the IN->OUT copy happens only if multiple processors exist. It means that
whether the response message is returned as OUT or IN depends on if the
route results in single processor or not. I think it looks a bit confusing
and wondering if we can just wrap it with the Pipleline even if there's
single processor. Is there any undesirable side effect to doing that?
Thanks,
Tomo
