James,
We assume that the class is loaded when ActiveMQObjectMessage.getObject() is
called, and the de-serialization takes place. Is this correct? Now, after
looking at the code, it seems that the class loader is resolved in
ClassLoadingAwareObjectInputStream. Is this correct? What can cause it t
Thanks, Christopher,
No, it's not. The broker and the consumer are on two different machines. The
transport is TCP.
We use Jencks for the consumer.
We're currently on ActiveMQ 4.1, but we had this problem on 4.0 as well.
Naaman
Christopher G. Stach II wrote:
>
> nlif wrote:
>> Thanks, James.
nlif wrote:
> Thanks, James.
>
> Yes, I'm also inclined to think it's a class loader issue. But then, how
> come it doesn't happen all the time?
> What may cause it to load another version of the class only some of the
> times?
> Does ActiveMQ do anything... um, unconventional with class-loading?
Thanks, James.
Yes, I'm also inclined to think it's a class loader issue. But then, how
come it doesn't happen all the time?
What may cause it to load another version of the class only some of the
times?
Does ActiveMQ do anything... um, unconventional with class-loading?
Thanks,
Naaman
James.S
If you get a ClassCastException yet the object seems to be of the
correct class name, then it looks like a class loader issue (multiple
versions of the same class on the classpath). Otherwise if it was a
serialization issue, you wouldn't get an object you can inspect.
On 4/30/07, nlif <[EMAIL PR
Hi all,
We have a strange problem, maybe someone has seen this or has an idea:
In our MessageListener, we cast the Message into ObejctMessage, and then
call getObject() to get the actual object we sent. Upon trying to cast the
object to the proper class (SendMobileNotification), we get a
ClassCa
